Upload
vanque
View
215
Download
1
Embed Size (px)
Citation preview
D i r e c t S t r e n g t h D e s i g n o f C o l d - F o r m e d S t e e l M e m b e r s w i t h P e r f o r a t i o n s
R E S E A R C H R E P O R T R P 0 9 - 1 M A R C H 2 0 0 9
Committee on Specif ications
for the Design of Cold-Formed
Steel Structural Members
re
sear
ch re
port
American Iron and Steel Institute
The material contained herein has been developed by researchers based on their research findings. The material has also been reviewed by the American Iron and Steel Institute Committee on Specifications for the Design of Cold-Formed Steel Structural Members. The Committee acknowledges and is grateful for the contributions of such researchers.
The material herein is for general information only. The information in it should not be used without first securing competent advice with respect to its suitability for any given application. The publication of the information is not intended as a representation or warranty on the part of the American Iron and Steel Institute, or of any other person named herein, that the information is suitable for any general or particular use or of freedom from infringement of any patent or patents. Anyone making use of the information assumes all liability arising from such use.
Copyright 2009 American Iron and Steel Institute
The Johns Hopkins University Department of Civil Engineering Latrobe Hall 210 Baltimore, MD 21218
Research Report
DIRECT STRENGTH DESIGN OF COLD‐FORMED STEEL MEMBERS
WITH PERFORATIONS by
Cristopher D. Moen Graduate Research Assistant B.W. Schafer Associate Professor Submitted to: American Iron and Steel Institute Committee on Specifications for the Design of Cold‐Formed Steel Structural Members 1140 Connecticut Ave, Suite 705 Washington, DC 20036
March 2009
ii
Abstract
Cold‐formed steel (CFS) structural members are commonly manufactured with holes
to accommodate plumbing, electrical, and heating conduits in the walls and ceilings of
buildings. Current design methods available to engineers for predicting the strength of
CFS members with holes are prescriptive and limited to specific perforation locations,
spacings, and sizes. The Direct Strength Method (DSM), a relatively new design method
for CFS members validated for members without holes, predicts the ultimate strength of
a general CFS column or beam with the elastic buckling properties of the member cross‐
section (e.g., plate buckling) and the Euler buckling load (e.g., flexural buckling). This
research project, sponsored by the American Iron and Steel Institute, extends the
appealing generality of DSM to cold‐formed steel beams and columns with perforations.
The elastic buckling properties of rectangular plates and cold‐formed steel beams
and columns, including the presence of holes, are studied with thin shell finite element
eigenbuckling analysis. Buckled mode shapes unique to members with holes are
categorized. Parameter studies demonstrate that critical elastic buckling loads either
decrease or increase with the presence of holes, depending on the member geometry and
hole size, spacing, and location. Simplified alternatives to FE elastic buckling analysis
for members with holes are developed with classical plate stability equations and freely
available finite strip analysis software.
iii
Experiments on cold‐formed steel columns with holes are conducted to observe the
interaction between elastic buckling, load–deformation response, and ultimate strength.
The experimental results are used to validate an ABAQUS nonlinear finite element
protocol, which is implemented to simulate loading to collapse of several hundred cold‐
formed steel beams and columns with holes. The results from these simulations,
supplemented with existing beam and column data, guide the development of design
equations relating elastic buckling and ultimate strength for cold‐formed steel members
with holes. These equations and the simplified elastic buckling prediction methods are
presented as a proposed design procedure for an upcoming revision to the American
Iron and Steel Institute’s North American Specification for the Design of Cold‐Formed
Steel Structural Members.
iv
Summary of Progress The primary goal of this AISI‐funded research is to extend the Direct Strength Method to cold‐formed steel members with holes. Research begins September 2005 Progress Report #1 February 2006 Accomplishments: • Evaluated the ABAQUS S9R5, S4, and S4R thin shell elements for accuracy and versatility in thin‐walled modeling problems • Studied the influence of element aspect ratio and element quantity when modeling rounded corners in ABAQUS • Developed custom MATLAB tools for meshing holes, plates, and cold‐formed steel members in ABAQUS • Determined the influence of a slotted hole on the elastic buckling of a structural stud channel and classified local, distortional, and global buckling modes • Investigated the influence of hole size on the elastic buckling of a structural stud channel • Performed a preliminary comparison of existing experimental data on cold‐formed steel columns with holes to DSM predictions • Conducted a study on the influence of the hole width to plate width ratio on the elastic buckling behavior of a simply supported rectangular plate Papers from this research: Moen, C.D., Schafer, B.W. (2006) “Impact of Holes on the Elastic Buckling of Cold‐Formed Steel Columns with Application to the Direct Strength Method.” Eighteenth International Specialty Conference on Cold‐Formed Steel Structures, Orlando, FL. Moen, C.D., Schafer, B.W. (2006) “Stability of Cold‐Formed Steel Columns With Holes.” Stability and Ductility of Steel Structures Conference, Lisbon, Portugal. Progress Report #2 August 2006 Accomplishments: • Evaluated the influence of slotted hole spacing on the elastic buckling of plates (with • Determined the impact of flange holes on the elastic buckling of an SSMA structural stud
v
• Conducted a preliminary investigation into the nonlinear solution algorithms available in ABAQUS • Compared the ultimate strength and load‐displacement response of a rectangular plate and an SSMA structural stud column with and without a slotted hole using nonlinear finite element models in ABAQUS • Calculated the effective width of a rectangular plate with and without a slotted hole using nonlinear finite element models in ABAQUS Progress Report #3 February 2007 Accomplishments: • Conducted an experimental study to evaluate the influence of a slotted web holes on the compressive strength, ductility, and failure modes of short and intermediate length C‐section channel columns • Studied the influence of slotted web holes on the elastic buckling behavior of cold‐formed steel C‐section beams and identified unique hole modes similar to those observed in compression members • Demonstrated that the Direct Strength Method is a viable predictor of ultimate strength for beams with holes Papers from this research: Moen, C.D., Schafer, B.W. (2008). “Experiments on cold‐formed steel columns with holes.” Thin‐Walled Structures, 46(10), 1164‐1182. Moen, C.D., Schafer, B.W. (2008). “Observing and quantifying the elastic buckling and tested response of cold‐formed steel columns with holes.” Fifth International Conference on Thin‐Walled Structures, Brisbane, Australia. Progress Report #4 July 2007 Accomplishments: • Completed experimental program including tensile coupon tests, elastic buckling study of specimens, and DSM strength comparison • Developed nonlinear finite element approach including a prediction method for residual stresses and completed preliminary finite element studies of the 24 column specimens • Proposed DSM approach for columns with holes, compared the options against the column database, conducted preliminary simulations to explore column strength curves
vi
Papers from this research: Moen, C.D., Igusa, T., Schafer, B.W. (2008). “Prediction of residual stresses and strains in cold‐formed steel members.” Thin‐Walled Structures, 46(11), 1274‐1289. Moen, C.D., Igusa, T., Schafer, B.W. (2008). “A mechanics‐based prediction method for residual stresses and initial plastic strains in cold‐formed steel structural members.” Fifth Conference on Coupled Instabilities in Metal Structures, Sydney, Australia. Progress Report #5 February 2008 Accomplishments: • Developed critical elastic buckling stress equations for stiffened and unstiffened
elements with holes under uniaxial compression. • Implemented finite strip approximation method for predicting the critical elastic
local and distortional buckling load of cold‐formed steel members with holes. • Derived and tested a method for predicting the Euler buckling loads of cold‐formed
steel columns and beams with holes, including flexural and flexural‐torsional buckling of columns and lateral‐torsional buckling of beams.
Papers from this research: Moen, C.D., Schafer, B.W. (2008). “Simplified methods for predicting elastic buckling of cold‐formed steel structural members with holes.” 19th International Specialty Conference on Cold‐Formed Steel Structures, St. Louis, Missouri.
Final Report Fall 2008
Accomplishments: • Created simulated column and beam experiment database with nonlinear finite
element analysis in ABAQUS and CUFSM elastic buckling approximate methods • Used the simulation results and existing column and beam experiment results to
develop and validate proposed DSM Holes design equations for CFS columns and beams
vii
Table of Contents 0HChapter 1 Introduction ............................................................................................................................. 442H1 1HChapter 2 Thin-shell finite element modeling in ABAQUS ............................................................... 443H8
2H2.1 Comparison of ABAQUS thin-shell elements ............................................................................ 444H9 3H2.2 Modeling holes in ABAQUS ....................................................................................................... 445H16 4H2.3 Modeling Rounded Corners in ABAQUS ................................................................................. 446H18 5H2.4 Summary of modeling guidelines .............................................................................................. 447H21
6HChapter 3 Elastic buckling of cold-formed steel cross-sectional elements with holes .................. 448H22 7H3.1 Plate and hole dimensions .......................................................................................................... 449H23 8H3.2 Finite element modeling assumptions ...................................................................................... 450H25 9H3.3 Stiffened element in uniaxial compression ............................................................................... 451H25 10H3.4 Stiffened element in bending ...................................................................................................... 452H43 11H3.5 Unstiffened element in uniaxial compression .......................................................................... 453H58
12HChapter 4 Elastic buckling of cold-formed steel members with holes ............................................ 454H66 13H4.1 Finite element modeling assumptions ...................................................................................... 455H67 14H4.2 Elastic buckling of columns with holes ..................................................................................... 456H67 15H4.3 Elastic buckling of beams with holes ....................................................................................... 457H126
16HChapter 5 Experiments on cold-formed steel columns with holes ................................................ 458H159 17H5.1 Acknowledgements ................................................................................................................... 459H160 18H5.2 Testing Program ......................................................................................................................... 460H160 19H5.3 Elastic buckling calculations ..................................................................................................... 461H196 20H5.4 Experiment results ..................................................................................................................... 462H205
21HChapter 6 Predicting residual stresses and plastic strains in cold-formed steel members ......... 463H223 22H6.1 Stress-strain coordinate system and notation ........................................................................ 464H226 23H6.2 Prediction method assumptions .............................................................................................. 465H227 24H6.3 Derivation of the residual stress prediction method ............................................................. 466H230 25H6.4 Derivation of effective plastic strain prediction method ...................................................... 467H240 26H6.5 Employing the prediction method in practice: quantifying the coil radius influence ...... 468H244 27H6.6 Comparison of prediction method to measured residual stresses ...................................... 469H248 28H6.7 Discussion ................................................................................................................................... 470H254 29H6.8 Acknowledgements ................................................................................................................... 471H257
30HChapter 7 Nonlinear finite element modeling of cold-formed steel structural members ........... 472H258 31H7.1 Preliminary nonlinear FE studies ............................................................................................ 473H259 32H7.2 Nonlinear finite element modeling of columns with holes .................................................. 474H288
33HChapter 8 The Direct Strength Method for cold-formed steel members with holes ................... 475H313 34H8.1 DSM for columns with holes .................................................................................................... 476H314 35H8.2 DSM for laterally braced beams with holes ............................................................................ 477H374
36HChapter 9 Conclusions and proposed future work .......................................................................... 478H409 37H9.1 Conclusions ................................................................................................................................. 479H409 38H9.2 Future work ................................................................................................................................ 480H411
39HReferences .................................................................................................................................................. 481H415 40HAppendix A ABAQUS input file generator in Matlab ................................................................ 482H419 41HAppendix B ABAQUS element-based elastic buckling results .................................................. 483H425 42HAppendix C Derivation of elastic buckling coefficients for unstiffened elements .................. 484H435 43HAppendix D Elastic buckling prediction method of cross-sectional elements with holes ...... 485H441 44HAppendix E Derivation of global critical elastic buckling load for a column with holes....... 486H445 45HAppendix F Column experiment results ...................................................................................... 487H449 46HAppendix G Residual stresses– backstress for kinematic hardening implementation ........... 488H474 47HAppendix H Experiment true stress-strain curves ....................................................................... 489H476
viii
48HAppendix I Column experiment nonlinear FE simulation results ............................................... 490H497 49HAppendix J Contact simulation in ABAQUS ................................................................................... 491H521 50HAppendix K Simulated column experiments database ............................................................... 492H524 51HAppendix L Simulated beam experiment database .................................................................... 493H528 52HAppendix M Comparison of tested strengths to predicted strengths with AISI S100-07 ....... 494H531
ix
List of Figures
53HFigure 1.1 Perforations are provided in structural studs to accommodate utilities in the walls of buildings ................................................................................................................................... 495H2
54HFigure 1.2 Hole patterns in storage rack columns ................................................................................... 496H2 55HFigure 1.3 Column elastic buckling curve generated with CUFSM ..................................................... 497H4 56HFigure 1.4 DSM global buckling failure design curve and equations .................................................. 498H4 57HFigure 1.5 DSM local buckling failure design curve and equations ..................................................... 499H5 58HFigure 1.6 DSM distortional buckling failure design curve and equations ......................................... 500H5 59HFigure 2.1 ABAQUS S4\S4R shell element with four nodes and a linear shape function, ABAQUS
S9R5 shell element with nine nodes and a quadratic shape function ............................ 501H10 60HFigure 2.2 Buckled shape of a stiffened plate ........................................................................................ 502H10 61HFigure 2.3 Accuracy of ABAQUS S9R5, S4, and S4R elements for a stiffened element with varying
aspect ratios, 8:1 finite element aspect ratio for the S9R5 element, 4:1 element aspect ratio for the S4 and S4R elements ....................................................................................... 503H13
62HFigure 2.4 Accuracy of S4, S4R, and S9R5 elements as a function of the number of elements provided per buckled half-wavelength, stiffened element, square waves (k=4).......... 504H14
63HFigure 2.5 Buckled shape of an unstiffened element, m=1 shown ...................................................... 505H15 64HFigure 2.6 Accuracy of S9R5 elements as the number of finite elements provided along an
unstiffened element varies, L/h=4 ...................................................................................... 506H15 65HFigure 2.7. Finite element mesh and plate dimensions: slotted, rectangular, and circular holes .. 507H17 66HFigure 2.8 Hole discretization using S9R5 elements ............................................................................. 508H18 67HFigure 2.9 The critical elastic buckling stress converges to a constant magnitude when the S9R5
element aspect ratio a/b is between 0.5 and 2 and element corner angles are skewed ................................................................................................................................................. 509H18
68HFigure 2.10 ABAQUS S9R5 initial curvature limit requires at least five elements to model corner ................................................................................................................................................. 510H19
69HFigure 2.11 SSMA 600S162-68 C-section corner modeled with a) one S9R5 element, b) three S9R5 elements .................................................................................................................................. 511H20
70HFigure 2.12 The number of S9R5 corner elements has a minimal influence on the critical elastic buckling loads of an SSMA 600162-68 C-section column with L=48 in. ........................ 512H21
71HFigure 3.1 Stiffened and unstiffened elements in a lipped C-section ................................................. 513H23 72HFigure 3.2 Element and hole dimension definitions ............................................................................. 514H24 73HFigure 3.3 Definition of unstiffened strip “A” and “B” for a plate with holes. ................................. 515H24 74HFigure 3.4 Definition of neutral axis location for stiffened elements in bending. ............................. 516H24 75HFigure 3.5 ABAQUS boundary conditions and loading conditions for a stiffened element in
uniaxial compression ............................................................................................................ 517H26 76HFigure 3.6 Influence of a slotted hole on the elastic buckling stress of a simply supported
rectangular plate with varying length ............................................................................... 518H27 77HFigure 3.7 Comparison of buckled shape and displacement contours for a rectangular plate with
hhole/h=0.66 and L/Lhole =3, (a) with slotted hole and (b) without hole. Notice the change in length and quantity of buckled cells with the addition of a slotted hole. ... 519H28
78HFigure 3.8 Buckled shape of a simply supported plate (a) with a slotted hole and (b) without a hole. L=15Lhole , hhole/h=0.66. The slotted hole dampens buckling but does not significantly change the natural half-wavelength of the plate........................................ 520H28
79HFigure 3.9 (a) Slotted hole causes local buckling (hhole/h=0.26), compared to (b) buckled cells at the natural half-wavelength of the plate ............................................................................ 521H29
80HFigure 3.10 Definition of center-to-center dimension for the slotted holes ....................................... 522H30
x
81HFigure 3.11 Influence of slotted hole spacing on the elastic buckling load of a long simply supported rectangular plate ................................................................................................ 523H30
82HFigure 3.12 Comparison of buckled shapes for a long stiffened element (L=24 Lhole ) with a slotted hole spacing of S/Lhole=4 and hhole/h=0.66, 0.44, and 0.26. .................................................. 524H31
83HFigure 3.13 Variation in fcr with increasing hhole/h for a stiffened element correspond to buckling mode shapes (see Figure 3.12 for examples of plate buckling and unstiffened strip buckling mode shapes) ......................................................................................................... 525H33
84HFigure 3.14 Unstiffened strip elastic buckling stress conversion from the net to the gross section ................................................................................................................................................. 526H35
85HFigure 3.15 Accuracy of stiffened element prediction method as a function of hole spacing S to plate width h (a) without and (b) with the dimensional limits in Eq. (3.8) and Eq.(3.9) ................................................................................................................................................. 527H38
86HFigure 3.16 Accuracy of stiffened element prediction method as a function of hole spacing S to length of hole Lhole (a) without and (b) with the dimensional limits in Eq. (3.8) and Eq.(3.9) .................................................................................................................................... 528H38
87HFigure 3.17 Accuracy of the stiffened element prediction method as a function of hole width hhole to plate width h (a) without and (b) with the dimensional limits in Eq. (3.8) and Eq.(3.9) .................................................................................................................................... 529H38
88HFigure 3.18 For plates where the unstiffened strip is narrow compared to the plate width, plate buckling occurs between the holes. .................................................................................... 530H39
89HFigure 3.19 Plate buckling and unstiffened strip buckling may both exist for a plate with holes. These modes are predicted conservatively as unstiffened strip buckling. ................... 531H40
90HFigure 3.20 Accuracy of prediction method for stiffened elements with square or circular holes as a function of hole width hhole to plate width h. .................................................................. 532H40
91HFigure 3.21 Accuracy of the stiffened element elastic buckling prediction method as a function of unstiffened strip width hstrip versus plate width h for offset holes (a) without and (b) with the dimensional limits in Eq. (3.8) and Eq.(3.9) ....................................................... 533H41
92HFigure 3.22 Holes at the edge of a wide stiffened plate reduce the axial stiffness (and critical elastic buckling stress) but do not change the buckled shape. ....................................... 534H42
93HFigure 3.23 Accuracy of the stiffened element elastic buckling prediction method as a function of hole offset δhole versus plate width h for offset holes (a) without and (b) with the dimensional limits in Eq. (3.8), Eq.(3.9), and Eq. (3.11) .................................................... 535H43
94HFigure 3.24 Boundary and loading conditions for a stiffened element in bending .......................... 536H43 95HFigure 3.25 Stiffened plates loaded with a linear bending stress gradient exhibit buckling of the
unstiffened strip adjacent to the hole in the compression region of the plate. ............. 537H45 96HFigure 3.26 Influence of slotted holes on critical elastic buckling stress fcr of stiffened elements in
bending as a function of (a) hole size relative to plate width and (b) hole spacing as a function of hole length. ........................................................................................................ 538H46
97HFigure 3.27 Hole location influence on critical elastic buckling stress fcr for a stiffened plate in bending (Y=0.50h) (Buckled mode shapes corresponding to A, B, C, and D are provided in Figure 3.28.) ...................................................................................................... 539H47
98HFigure 3.28 The buckled mode shape changes as slotted holes move from the compression region to the tension region of a stiffened element in bending (hhole/h=0.20). ........................... 540H48
99HFigure 3.29 Hole location influence on critical elastic buckling stress fcr for a stiffened plate in bending (Y=0.75h) ................................................................................................................. 541H49
100HFigure 3.30 Derivation of stress ratio ψΑ for unstiffened strip “A”. ................................................... 542H51 101HFigure 3.31 Derivation ψB and conversion of the compressive stress at the edge of unstiffened
strip “B” to the stress fcrB at the edge of the plate ............................................................. 543H52 102HFigure 3.32 Derivation of fcrh for the case when hA+hhole≥Y (when the hole is located partially in
the compressed region and partially in the tension region of the plate) ....................... 544H53
xi
103HFigure 3.33 Derivation of fcrh for the case when hA+hhole<Y (hole lies completely in the compressed region of the plate). ............................................................................................................... 545H54
104HFigure 3.34 Influence of Lhole/yA on the accuracy of the prediction method for stiffened elements in bending (a) without and (b) with the dimensional limits defined in Eq. (3.9), Eq. (3.24), Eq. (3.25), and Eq. (3.26). ........................................................................................... 546H56
105HFigure 3.35 Influence of hA/Y on the accuracy of the prediction method for stiffened elements in bending (a) without and (b) with the dimensional limits defined in Eq. (3.9), Eq. (3.24), Eq. (3.25), and Eq. (3.26). ........................................................................................... 547H56
106HFigure 3.36 Influence of S/h on the accuracy of the prediction method for stiffened elements in bending (a) without and (b) with the dimensional limits defined in Eq. (3.9), Eq. (3.24), Eq. (3.25), and Eq. (3.26). ........................................................................................... 548H57
107HFigure 3.37 Influence of S/Lhole on the accuracy of the prediction method for stiffened elements in bending (a) without and (b) with the dimensional limits defined in Eq. (3.9), Eq. (3.24), Eq. (3.25), and Eq. (3.26). ........................................................................................... 549H57
108HFigure 3.38 Influence of h/hhole on the accuracy of the prediction method for stiffened elements in bending (a) without and (b) with the dimensional limits defined in Eq. (3.9), Eq. (3.24), Eq. (3.25), and Eq. (3.26). ........................................................................................... 550H58
109HFigure 3.39 ABAQUS boundary and loading conditions for unstiffened plate loaded uniaxially. ................................................................................................................................................. 551H58
110HFigure 3.40 The presence of holes causes a decrease in critical elastic buckling load for unstiffened plates in uniaxial compression. ...................................................................... 552H60
111HFigure 3.41 Buckled shapes of unstiffened plates with holes. ............................................................. 553H60 112HFigure 3.42 The critical elastic buckling stress of a stiffened plate decreases as holes are shifted
toward the simply supported edge (+δhole) ....................................................................... 554H62 113HFigure 3.43 The critical elastic buckling stress for stiffened elements with (a) transversely offset
holes and (b) centered holes (from Section 3.5.2) decreases as a function of hole length Lhole to hA ..................................................................................................................... 555H62
114HFigure 3.44 (a) Comparison of ABAQUS and empirical plate buckling coefficients for an unstiffened element with holes and (b) ABAQUS to predicted elastic buckling stress for an unstiffened element ................................................................................................... 556H64
115HFigure 4.1 C-section and hole dimension notation ............................................................................... 557H68 116HFigure 4.2 Columns are modeled with pinned warping-free boundary conditions and
compressed from both ends ................................................................................................ 558H69 117HFigure 4.3(a) SSMA 250S162-33 web plate and structural stud, and (b) SSMA 400S162-33 web
plate and structural stud ...................................................................................................... 559H71 118HFigure 4.4. Effect of a slotted hole on the elastic buckling load of simply supported plates and
structural studs ...................................................................................................................... 560H72 119HFigure 4.5 The presence of a hole creates unique local buckling modes where unstiffened strip
buckling adjacent to the hole occurs symmetrically (LH) or asymmetrically (LH2) increase the distortional tendency of the flanges.............................................................. 561H74
120HFigure 4.6 SSMA slotted hole location and local buckling LH mode, L=48 in., x/L=0.06,0.125,0.25,0.375,0.50. Note the distortional tendencies of the flanges at the hole. ......................................................................................................................................... 562H74
121HFigure 4.7 Influence of a slotted hole on the (a) distortional (D) and (b) global flexural-torsional (GFT) modes of a cold-formed steel column ..................................................................... 563H75
122HFigure 4.8 Influence of SSMA slotted hole location on Pcr for a 362S162-33 C-section (refer to Figure 4.5, Figure 4.6, and Figure 4.7 for buckled shape summaries) ........................... 564H76
123HFigure 4.9 Connection detail for structural stud to exterior wall requires a screw or bolt hole placed in the stud flange (Western States Clay Products Association 2004) ................ 565H77
124HFigure 4.10 Influence of flange hole diameter on the local (L), distortional (D), and global (GFT) elastic buckling loads of an SSMA 362S162-33 structural stud ....................................... 566H78
xii
125HFigure 4.11 Local (L) buckling is dominated by flange and web deformation near the holes as bhole/b exceeds 0.70 .................................................................................................................. 567H78
126HFigure 4.12 Experimental program boundary conditions as implemented in ABAQUS ................ 568H80 127HFigure 4.13 Influence of fixed-fixed boundary conditions versus warping free boundary
conditions on Pcrd for column experiments(L/H<4 ) as a function of (a) column length to fundamental distortional half-wavelength calculated with CUFSM and (b) column length to member length. ..................................................................................................... 569H87
128HFigure 4.14 Influence of fixed-fixed boundary conditions versus warping free boundary conditions on Pcrl for column experiments ( L/H<4) as a function of (a) hole width relative to column width and (b) hole length relative to column length ...................... 570H87
129HFigure 4.15 Influence of weak-axis pinned boundary conditions versus warping free boundary conditions on (a) Pcrl as a function of hole length to column length and (b) Pcrd as a function of column length to member length. .................................................................. 571H88
130HFigure 4.16 Rules for modeling a column net cross-section in CUFSM ............................................. 572H91 131HFigure 4.17 Local elastic buckling curve of net cross-section when (a) hole length is less than Lcrh
and (b) when hole length is greater than Lcrh..................................................................... 573H92 132HFigure 4.18 Comparison of CUFSM and ABAQUS predictions of unstiffened strip buckling. ...... 574H93 133HFigure 4.19 ABAQUS results verify CUFSM local buckling predictions for an SSMA 362S162-33
column with evenly spaced web holes. ............................................................................. 575H94 134HFigure 4.20 CUFSM and ABAQUS local buckling mode shapes are consistent when considering a
slotted flange hole. ................................................................................................................ 576H95 135HFigure 4.21 ABAQUS results verify CUFSM predictions for an SSMA 362S162-33 cross section
with evenly spaced flange holes. ........................................................................................ 577H95 136HFigure 4.22 ABAQUS predicts local plate buckling with distortional buckling interaction which is
not detected in CUFSM. ....................................................................................................... 578H96 137HFigure 4.23 ABAQUS results are slightly lower than CUFSM predictions, CUFSM predicts
correctly that plate local buckling controls over unstiffened strip buckling. ............... 579H97 138HFigure 4.24 Predicted Pcrh (CUFSM, buckling of the net cross-section) and Pcr (CUFSM, buckling
of the gross cross section, no hole) are compared relative to the ABAQUS Pcrl with experiment boundary conditions as a function of (a) hole width to flat web width and (b) hole length to column length ......................................................................................... 580H98
139HFigure 4.25 Predicted Pcrl (CUFSM approximate method) is compared relative to the ABAQUS Pcrl with experiment boundary conditions as a function of (a) hole width to flat web width and (b) hole length to column length ...................................................................... 581H99
140HFigure 4.26 CUFSM approximate method for calculating Pcrd for a column with holes. ............... 582H101 141HFigure 4.27 Modified cross section to be used in CUFSM to predict Pcrd for a column with holes.
............................................................................................................................................... 583H102 142HFigure 4.28 ABAQUS boundary conditions and imposed rotations for web plate ........................ 584H103 143HFigure 4.29 Plate deformation from imposed edge rotations, hhole/h=0.50 ..................................... 585H104 144HFigure 4.30 Transverse rotational stiffness of the plate is significantly reduced in the vicinity of
the slotted hole .................................................................................................................... 586H104 145HFigure 4.31 Comparison of CUFSM and ABAQUS distortional buckling mode shapes. .............. 587H107 146HFigure 4.32 CUFSM distortional buckling prediction method is conservative when considering
an SSMA 262S162-68 column with uniformly spaced holes. ........................................ 588H107 147HFigure 4.33 Warping-fixed boundary condition amplification of Pcrd .............................................. 589H109 148HFigure 4.34 Accuracy of the CUFSM approximate method for predicting Pcrd improves as column
length increases relative to the fundamental distortional half-wavelength for warping-fixed columns ...................................................................................................... 590H110
149HFigure 4.35 A “weighted thickness” cross section can be input directly into a program that solves the classical cubic stability equation for columns (e.g. CUTWP). ................................ 591H113
xiii
150HFigure 4.36 Weak-axis flexural and flexural-torsional global buckling modes for an SSMA 1200S162-68 column with evenly spaced circular holes. ............................................... 592H114
151HFigure 4.37 Variation in net section properties as circular hole diameter increases. ..................... 593H115 152HFigure 4.38 Comparison of “weighted thickness” and “weighted properties” cross-sectional area.
............................................................................................................................................... 594H116 153HFigure 4.39 Comparison of “weighted thickness” and “weighted properties” strong axis moment
of inertia. .............................................................................................................................. 595H117 154HFigure 4.40 Comparison of “weighted thickness” and “weighted properties” weak axis moment
of inertia. .............................................................................................................................. 596H117 155HFigure 4.41 ABAQUS boundary conditions for warping free and applied unit twist at x=0 in. and
warping free but rotation restrained at x=100 in. ........................................................... 597H119 156HFigure 4.42 Angle of twist decreases linearly in the SSMA 1200S162-68 column with warping free
end conditions. .................................................................................................................... 598H119 157HFigure 4.43 The “weighted properties” approximation for Javg matches closely with the ABAQUS
prediction for the SSMA 12S00162-68 column with holes ............................................. 599H120 158HFigure 4.44 ABAQUS boundary conditions for warping free and applied unit twist at x=0 in. and
warping fixed and rotation restrained at x=100 in. ........................................................ 600H121 159HFigure 4.45 Angle of twist is nonlinear along the SSMA 1200S162-68 column with warping fixed
end conditions at x=100 in. ................................................................................................ 601H122 160HFigure 4.46 Comparison of “weighted thickness” and “weighted properties” approximations to
the ABAQUS derived warping torsion constant Cw,avg. ................................................ 602H123 161HFigure 4.47 Comparison of “weighted thickness” and “weighted properties” prediction methods
for the SSMA 1200S162-68 weak-axis flexural buckling mode. Predictions using net section properties are also plotted as a conservative benchmark. ............................... 603H124
162HFigure 4.48 Comparison of “weighted thickness” and “weighted properties” prediction methods for the SSMA 1200S162-68 flexural-torsional column buckling mode. Predictions using net section properties are also plotted as a conservative benchmark. .............. 604H125
163HFigure 4.49 Cross section of beam specimen showing aluminum strap angles connected to C- flanges ................................................................................................................................... 605H127
164HFigure 4.50 C-section and hole dimension notation ........................................................................... 606H127 165HFigure 4.51 Experiment test setup with hole spacing, location of lateral bracing, spacing of
aluminum angle straps, and load points ......................................................................... 607H129 166HFigure 4.52 Finite element model boundary conditions for beam eigenbuckling analyses .......... 608H130 167HFigure 4.53 Channel and hole meshing details and modeling of aluminum angle straps ............ 609H131 168HFigure 4.54 ABAQUS meshing details for C-section rounded corners ............................................ 610H132 169HFigure 4.55 Modeling of the beam concentrated loads in ABAQUS ................................................ 611H133 170HFigure 4.56 Local buckling modes for specimen 2B,20,1&2(H) with and without holes .............. 612H135 171HFigure 4.57 Local buckling modes for specimen 3B,14,1&2(H) with and without holes .............. 613H135 172HFigure 4.58 Local buckling modes for specimen 6B,18,1&2(H) with and without holes ............. 614H136 173HFigure 4.59 Local buckling modes for specimen BP-40(H) with and without holes ..................... 615H136 174HFigure 4.60 Local buckling modes for specimen 12B,16,1&2(H) with and without holes ............ 616H137 175HFigure 4.61 Distortional buckling modes for specimen 2B,20,1&2(H) with and without holes .. 617H138 176HFigure 4.62 Distortional buckling modes for specimen 3B,14,1&2(H) with and without holes .. 618H139 177HFigure 4.63 Distortional buckling modes for specimen 6B,18,1&2(H) with and without holes .. 619H139 178HFigure 4.64 Distortional buckling modes for specimen BP5-40(H) with and without holes ......... 620H140 179HFigure 4.65 Distortional buckling modes for specimen 12B,16,1&2(H) with and without holes 621H140 180HFigure 4.66 Elastic buckling curve for 12” deep specimen with modal participation summarized,
note that selected L and D are mixed local-distortional modes ................................... 622H141 181HFigure 4.67 Possible global buckling mode occurs about the compression flange lateral brace
point ...................................................................................................................................... 623H142
xiv
182HFigure 4.68 Influence of holes on beam specimen Mcrl (Channel 1 and Channel 2 plotted) considering (a) all local buckling modes and (b) the lowest local buckling mode .... 624H147
183HFigure 4.69 Influence of holes on beam specimen Mcrd (Channel 1 and Channel 2 plotted) as a function of hole depth to flat web depth considering (a) all distortional buckling modes and (b) the lowest distortional buckling mode .................................................. 625H148
184HFigure 4.70 Influence of holes on beam specimen Mcrd (Channel 1 and Channel 2 plotted) as a function of web depth to flange width considering (a) all distortional buckling modes and (b) the lowest distortional buckling mode ............................................................... 626H148
185HFigure 4.71 Influence of test boundary conditions on Mcrl ................................................................ 627H149 186HFigure 4.72 Influence of test boundary conditions on (a) Mcrd and (b) on the distortional half-
wavelength ........................................................................................................................... 628H151 187HFigure 4.73 Boost in Mcrd from the angle restraints increases as the fundamental distortional half-
wavelength increases relative to the restraint spacing Sbrace .......................................... 629H151 188HFigure 4.74 Guidelines for restraining beam net cross-sections in the CUFSM local buckling
approximate method .......................................................................................................... 630H154 189HFigure 4.75 Comparison of ABAQUS to predicted Mcrl for C-sections with holes in the beam
database as a function of (a) web depth and (b) hole width relative to flat web depth ............................................................................................................................................... 631H155
190HFigure 4.76 Comparison of “mechanics-based” and “weighted-average” prediction methods to ABAQUS results for the distortional buckling load Mcrd of C-sections with holes in the elastic buckling database ............................................................................................. 632H156
191HFigure 4.77 ABAQUS boundary conditions and applied loading for an SSMA 1200S162-68 beam with holes (hhole/h=0.50 shown) ........................................................................................ 633H157
192HFigure 4.78 Comparison of “weighted thickness” and “weighted properties” prediction methods for the SSMA 1200S162-68 lateral-torsional beam buckling mode. Predictions using net section properties are also plotted as a conservative benchmark. ......................... 634H158
193HFigure 5.1 Column testing parameters and naming convention ....................................................... 635H161 194HFigure 5.2 Tested lengths of cold-formed steel columns with holes as a function of (a) column
length L and and (b) L versus out-to-out column width H ........................................... 636H163 195HFigure 5.3 Typical column specimens with slotted holes .................................................................. 637H164 196HFigure 5.4 Column test setup and instrumentation ............................................................................ 638H165 197HFigure 5.5 Novotechnik position transducer with ball-jointed magnetic tip .................................. 639H166 198HFigure 5.6 Central Machinery metal band saw used to rough cut column specimens .................. 640H167 199HFigure 5.7 362S162-33 short column specimen with bismuth end diaphragms .............................. 641H168 200HFigure 5.8 600S162-33 short column specimen oriented in CNC machine ...................................... 642H169 201HFigure 5.9 An end mill is used to prepare the column specimens .................................................... 643H169 202HFigure 5.10 The intermediate length specimens were end-milled in a manual milling machine . 644H170 203HFigure 5.11 The specimens are clamped at the webs only to avoid distortion of the cross-section
............................................................................................................................................... 645H170 204HFigure 5.12 Setup procedure for measuring specimen cross section dimensions .......................... 646H172 205HFigure 5.13 Procedure for measuring specimen cross-section dimensions ..................................... 647H173 206HFigure 5.14 Procedure for measuring flange-lip and flange-web angles ......................................... 648H174 207HFigure 5.15 Specimen measurement nomenclature ............................................................................ 649H175 208HFigure 5.16 Base metal and zinc thickness definitions ....................................................................... 650H177 209HFigure 5.17 Removal of tensile coupon zinc coating as a function of time ...................................... 651H179 210HFigure 5.18 A height gauge is used to measure specimen length ..................................................... 652H180 211HFigure 5.19 Lengths are measured at the four corners of the C-section column ............................. 653H181 212HFigure 5.20 A dial gauge and precision stand are used to measure initial web imperfections .... 654H184 213HFigure 5.21 Web imperfection measurement grid and coordinate system ...................................... 655H184 214HFigure 5.22 Column specimen alignment schematic .......................................................................... 656H186
xv
215HFigure 5.23 Influence of platen bending stiffness on end moments for a fixed-fixed eccentric column .................................................................................................................................. 657H188
216HFigure 5.24 Column specimen weak axis out-of-straightness schematic ......................................... 658H189 217HFigure 5.25 Digital calipers are used to measure the distance from the column web to platen edge
............................................................................................................................................... 659H190 218HFigure 5.26 Tensile coupons are first rough cut with a metal ban saw ............................................ 660H192 219HFigure 5.27 Tensile coupon dimensions as entered in the CNC milling machine computer ........ 661H192 220HFigure 5.28 A custom jig allows three tensile coupons to be milled at once in the CNC machine
............................................................................................................................................... 662H193 221HFigure 5.29 ATS machine used to test tensile coupons....................................................................... 663H194 222HFigure 5.30 Gradually yielding stress-strain curve with 0.2% strain offset method ...................... 664H195 223HFigure 5.31 Sharp-yielding stress strain curve using an autographic method for determining Fy
............................................................................................................................................... 665H195 224HFigure 5.32 (a) Local and distortional elastic buckled mode shapes for (a) short (L=48 in.) 362S162-
33 specimens and (b) intermediate length (L=48 in.) 362S162-33 specimens. ............. 666H198 225HFigure 5.33 Local and distortional elastic buckled mode shapes for (a) short (L=48 in.) 600S162-33
specimens and (b) intermediate length (L=48 in.) 600S162-33 specimens. .................. 667H198 226HFigure 5.34 Local (L) and distortional (D) DSM strength predictions are similar in magnitude for
both 362S162-33 and 600S162-33 cross-sections, indicating that L-D modal interaction will occur during the tested response of the columns. .................................................. 668H201
227HFigure 5.35 Comparison of global mode shapes for intermediate length 362S162-33 and 600S162-33 specimens. ....................................................................................................................... 669H205
228HFigure 5.36 Load-displacement progression for short column specimen 362-2-24-NH ................ 670H207 229HFigure 5.37 Load-displacement progression for short column specimen 362-2-24-H ..................... 671H208 230HFigure 5.38 Load-displacement curve for a 362S162-33 short column with, without a slotted hole
............................................................................................................................................... 672H208 231HFigure 5.39 Comparison of load-deformation response and lateral flange displacements for
specimen 362-2-24-NH ....................................................................................................... 673H210 232HFigure 5.40 Influence of a slotted hole on 362S162-33 short column lateral flange displacement 674H210 233HFigure 5.41 Load-displacement progression for short column specimen 600-1-24-NH ................ 675H212 234HFigure 5.42 Load-displacement progression for short column specimen 600-1-24-H .................... 676H212 235HFigure 5.43 Comparison of load-displacement response for short 600S162-33 column specimens
with and without holes ...................................................................................................... 677H213 236HFigure 5.44 Load-displacement progression, intermediate length column specimen 362-3-48-NH
............................................................................................................................................... 678H215 237HFigure 5.45 Load-displacement progression for intermediate length column specimen 362-3-48-H
............................................................................................................................................... 679H215 238HFigure 5.46 Load-displacement curve, 362S162-33 intermediate column with and without a hole
............................................................................................................................................... 680H216 239HFigure 5.47 362S162-33 long column mid-height flange displacements show the global torsional
failure mode ......................................................................................................................... 681H216 240HFigure 5.48 Load-displacement progression, intermediate length column specimen 600-1-48-NH
............................................................................................................................................... 682H218 241HFigure 5.49 Load-displacement progression, intermediate length column specimen 600-1-48-NH
............................................................................................................................................... 683H218 242HFigure 5.50 Load-displacement comparison of intermediate length 600S162-33 specimens with
and without holes ............................................................................................................... 684H219 243HFigure 5.51 Short 600S162-33 column flange-lip corner lifts off platen during post-peak portion of
test ......................................................................................................................................... 685H222 244HFigure 6.1 Cold-formed steel roll-forming: (left) Sheet coil enters roll-forming line, (right) steel
sheet is cold-formed into C-shape cross-section (photos courtesy of Bradbury Group). ............................................................................................................................................... 686H224
xvi
245HFigure 6.2 Forming a bend: plastic bending and elastic springback of thin sheets results in a nonlinear through-thickness residual stress distribution. ............................................. 687H225
246HFigure 6.3 Stress-strain coordinate system as related to the coiling and cold-forming processes. ............................................................................................................................................... 688H226
247HFigure 6.4 Roll-forming setup with sheet coil fed from the (a) top of the coil and (b) bottom of coil. The orientation of the coil with reference to the roll-forming bed influences the direction of the coiling residual stresses. ......................................................................... 689H229
248HFigure 6.5 Coiling of the steel sheet may result in residual curvature which results in bending residual stresses as the sheet is flattened. ........................................................................ 690H231
249HFigure 6.6 Longitudinal residual stress distribution from coiling. ................................................... 691H232 250HFigure 6.7 Predicted longitudinal residual stress distribution from coiling, uncoiling, and
flattening of a steel sheet. ................................................................................................... 692H233 251HFigure 6.8 Cold-forming of a steel sheet. .............................................................................................. 693H235 252HFigure 6.9 Fully plastic transverse stress state from cold-forming. .................................................. 694H235 253HFigure 6.10 Force couple (Fp·½t) applied to simulate the elastic springback of the steel sheet after
the imposed radial deformation is removed. .................................................................. 695H236 254HFigure 6.11 Cold-forming of a steel sheet occurs as plastic bending and elastic springback,
resulting in a self-equilibrating transverse residual stress. ........................................... 696H237 255HFigure 6.12 Plastic bending and elastic springback from cold-forming in the transverse direction
result in longitudinal residual stresses because of the plane strain conditions. ......... 697H238 256HFigure 6.13 Flowchart summarizing the prediction method for residual stresses in roll-formed
members. .............................................................................................................................. 698H239 257HFigure 6.14 Plastic strain distribution from sheet coiling with a radius less than elastic-plastic
threshold rep. ......................................................................................................................... 699H241 258HFigure 6.15 Effective plastic strain in a cold-formed steel member from sheet coiling when the
radius rx is less than the elastic-plastic threshold rep. ...................................................... 700H242 259HFigure 6.16 Effective von Mises true plastic strain at the location of cold-forming of a steel sheet.
............................................................................................................................................... 701H243 260HFigure 6.17 Flowchart summarizing the prediction method for effective plastic strains in roll-
formed members ................................................................................................................. 702H244 261HFigure 6.18 Coil coordinate system and notation. ............................................................................... 703H245 262HFigure 6.19 Influence of sheet thickness and yield stress on through-thickness longitudinal
residual stresses (z-direction, solid lines are predictions for mean coil radius, dashed lines for mean radius +/- one standard deviation). ....................................................... 704H247
263HFigure 6.20 The mean-squared error of the predicted and measured bending residual stresses for de M. Batista and Rodrigues (De Batista and Rodrigues 1992), Specimen CP1 is minimized when rx=1.60rinner. ............................................................................................ 705H251
264HFigure 6.21 (a) Histogram and (b) scattergram of bending residual stress prediction error (flat cross-sectional elements) for 18 roll-formed specimens. ............................................... 706H254
265HFigure 6.22 (a) Histogram and (b) scattergram of bending residual stress prediction error (corner cross-sectional elements) for 18 roll-formed specimens. ............................................... 707H254
266HFigure 6.23 Definition of apparent yield stress, effective residual stress, and effective plastic strain as related to a uniaxial tensile coupon test. .......................................................... 708H256
267HFigure 7.1 True stress-strain curve derived from a tensile coupon test (Yu 2005) ........................... 709H260 268HFigure 7.2 Simply supported boundary conditions with equation constraint coupling at loaded
edges ..................................................................................................................................... 710H261 269HFigure 7.3 Application of (a) uniform load and (b) uniform displacement to a stiffened element
............................................................................................................................................... 711H262 270HFigure 7.4 Type 1 imperfection (Schafer and Peköz 1998) ................................................................. 712H263 271HFigure 7.5 Modified Riks method load-displacement solutions and failure modes ...................... 713H265 272HFigure 7.6 Correlation between initial imperfection shape and fold line locations at failure ....... 714H265
xvii
273HFigure 7.7 Artificial damping load-displacement solutions and failure modes ............................. 715H268 274HFigure 7.8 Stiffened element boundary conditions with rigid body coupling at loaded edges ... 716H269 275HFigure 7.9 Initial geometric imperfection field used for the stiffened element with and without a
hole ........................................................................................................................................ 717H270 276HFigure 7.10 Deformation at ultimate load of a stiffened element with a hole loaded in
compression. The common failure mechanism is material yielding adjacent to the hole followed by plate folding. ......................................................................................... 718H271
277HFigure 7.11 Load-displacement curve for the RIKS1 and RIKS2 models showing direction reversal along load path ..................................................................................................... 719H272
278HFigure 7.12 RIKS1 and RIKS2 models experience convergence problems and return along the loading path, the RIKS3 model successfully predicts a peak load and finds a post-peak load path ..................................................................................................................... 720H274
279HFigure 7.13 STATIC1 and STATIC2 load-displacement curves demonstrate convergence difficulties near the peak load. .......................................................................................... 721H275
280HFigure 7.14 STAB1 and STAB2 load-displacement curves demonstrate a highly nonlinear post-peak equilibrium path ........................................................................................................ 722H277
281HFigure 7.15 The STAB1 and STAB2 models (artificial damping, displacement control) exhibit a sharp drop in load as folding of the plate initiates near the hole. The STAB3 model (artificial damping, load control) finds the compressive load at which a complete loss of stiffness occurs. ............................................................................................................... 723H279
282HFigure 7.16 Comparison of ultimate limit state and elastic buckling plate behavior, initial imperfections are not considered in these results .......................................................... 724H280
283HFigure 7.17 Load-displacement sensitivity to imperfection magnitude for a plate without a hole ............................................................................................................................................... 725H282
284HFigure 7.18 Load-displacement sensitivity to imperfection magnitude for a plate with a slotted hole ........................................................................................................................................ 726H282
285HFigure 7.19 Calculation of “effective width” at a cross-section along a stiffened element ............ 727H284 286HFigure 7.20 Definition of longitudinal (S11) membrane stress .......................................................... 728H284 287HFigure 7.21 (a) longitudinal membrane stresses and (b) effective width of a stiffened element at
failure .................................................................................................................................... 729H285 288HFigure 7.22 (a) longitudinal membrane stresses and (b) effective width of a stiffened element with
a slotted hole at failure ....................................................................................................... 730H286 289HFigure 7.23 Effective width comparison for a plate with and without a slotted hole .................... 731H286 290HFigure 7.24 Through the thickness variation of effective width of a plate without a hole ............ 732H287 291HFigure 7.25 Through the thickness variation of effective width of a plate with a slotted hole ..... 733H287 292HFigure 7.26 Through thickness variation in longitudinal (S11) stresses in a plate at failure ......... 734H288 293HFigure 7.27 ABAQUS boundary conditions simulating column experiments ................................ 735H290 294HFigure 7.28 ABAQUS plastic strain curve for specimen 362-1-24-NH assuming (a) plasticity
initiates at the proportional limit and (b) plasticity initiates at 0.2% offset yield stress ............................................................................................................................................... 736H293
295HFigure 7.29 ABAQUS plastic strain curve for specimen 600-1-24-NH assuming (a) plasticity initiates at the proportional limit and (b) plasticity initiates at the beginning of the yield plateau (refer to Appendix H for the details on the development of this curve). ............................................................................................................................................... 737H293
296HFigure 7.30 Influence of ABAQUS material model on the load-deformation response of specimen 600-1-24-NH (work this figure with Figure 7.29) ........................................................... 738H294
297HFigure 7.31 Slotted holes are filled with S9R5 elements to obtain no hole imperfection shapes .. 739H295 298HFigure 7.32 L and D imperfection magnitudes described with a CDF (Schafer and Peköz 1998) 740H297 299HFigure 7.33 Method for measuring distortional imperfection magnitudes from experiments ..... 741H297 300HFigure 7.34 Definition of out-of-straightness imperfections implemented in ABAQUS ............... 742H298 301HFigure 7.35 ABAQUS element local coordinate system for use with residual stress definitions . 743H300 302HFigure 7.36 Transverse residual stress distribution applied at the corners of the cross-section ... 744H300
xviii
303HFigure 7.37 Longitudinal residual stress distribution applied at the corners of the cross-section ............................................................................................................................................... 745H300
304HFigure 7.38 Equivalent plastic strain distribution at the corners of the cross-section .................... 746H301 305HFigure 7.39 Influence of section points on the unbalanced moment (accuracy) of the transverse
residual stress distribution as implemented in ABAQUS ............................................. 747H302 306HFigure 7.40 Load-displacement response of specimen 362-1-24-NH ............................................... 748H306 307HFigure 7.41 Load-displacement response of specimen 362-1-24-H ................................................... 749H306 308HFigure 7.42 Load-displacement response of specimen 362-1-48-NH ............................................... 750H307 309HFigure 7.43 Load-displacement response of specimen 362-1-48-H ................................................... 751H307 310HFigure 7.44 Load-displacement response of specimen 600-1-24-NH ............................................... 752H308 311HFigure 7.45 Load-displacement response of specimen 600-2-24-H ................................................... 753H308 312HFigure 7.46 Load-displacement response of specimen 600-1-48-NH ............................................... 754H309 313HFigure 7.47 Load-displacement response of specimen 600-3-48-H ................................................... 755H309 314HFigure 7.48 Influence of residual stresses (RS) and plastic strains (PS) on the FE load-
displacement response of specimen 600-1-24-NH .......................................................... 756H311 315HFigure 7.49 Influence of residual stresses (RS) and plastic strains (PS) on the FE load-
displacement response of specimen 362-1-24-NH. ......................................................... 757H312 316HFigure 8.1 ABAQUS simulated column experiments boundary conditions and application of
loading .................................................................................................................................. 758H315 317HFigure 8.2 SSMA 800S250-97 structural stud with web holes considered in the DSM distortional
buckling study ..................................................................................................................... 759H318 318HFigure 8.3 SSMA 800S250-97 structural stud failure mode transition from distortional buckling to
yielding at the net section .................................................................................................. 760H320 319HFigure 8.4 Comparison of simulated column strengths (Anet/Ag=1.0) to (a) the existing DSM
distortional buckling design curve and to (b) the proposed DSM distortional buckling curve for columns with holes ............................................................................ 761H321
320HFigure 8.5 Comparison of simulated column strengths (Anet/Ag=0.90) to (a) the existing DSM distortional buckling design curve and to (b) the proposed DSM distortional buckling curve for columns with holes ............................................................................ 762H321
321HFigure 8.6 Comparison of simulated column strengths (Anet/Ag=0.80) to (a) the existing DSM distortional buckling design curve and to (b) the proposed DSM distortional buckling curve for columns with holes ............................................................................ 763H322
322HFigure 8.7 Comparison of simulated column strengths (Anet/Ag=0.70) to (a) the existing DSM distortional buckling design curve and to (b) the proposed DSM distortional buckling curve for columns with holes ............................................................................ 764H322
323HFigure 8.8 Comparison of simulated column strengths (Anet/Ag=0.60) to (a) the existing DSM distortional buckling design curve and to (b) the proposed DSM distortional buckling curve for columns with holes ............................................................................ 765H322
324HFigure 8.9 Comparison of simulated column strengths (Anet/Ag=1.00) to (a) the existing DSM global buckling design curve and to (b) the proposed DSM distortional buckling curve for columns with holes ............................................................................................ 766H325
325HFigure 8.10 Comparison of simulated column strengths (Anet/Ag=0.90) to (a) the existing DSM global buckling design curve and to (b) the proposed DSM distortional buckling curve for columns with holes ............................................................................................ 767H326
326HFigure 8.11 Comparison of simulated column strengths (Anet/Ag=0.80) to (a) the existing DSM global buckling design curve and to (b) the proposed DSM distortional buckling curve for columns with holes ............................................................................................ 768H326
327HFigure 8.12 Comparison of column test-to-prediction ratios for columns (Anet/Ag=1.0) failing by local-global buckling interaction as a function of (a) local slenderness (b) global slenderness ........................................................................................................................... 769H328
xix
328HFigure 8.13 Comparison of column test-to-prediction ratios for columns failing by local-global buckling interaction with Pne calculated (a) without the influence of holes (b) and with the influence of holes ................................................................................................. 770H330
329HFigure 8.14 Comparison of column test-to-prediction ratios for columns failing by local-global buckling interaction as a function of Pynet/Pne where Pne is calculated (a) without the influence of holes (b) and with the influence of holes ................................................... 771H331
330HFigure 8.15 SSMA 350S162-68 column failure mode changes from distortional-flexural torsional buckling failure to weak axis flexure as hole size increases (L=34 in.) ........................ 772H331
331HFigure 8.16 SSMA 800S250-43 (L=74 in.) column web local buckling changes to unstiffened strip buckling at peak load as hole size increases .................................................................... 773H332
332HFigure 8.17 Comparison of DSM local buckling design curve options when Pynet=0.80 Pyg and (a) Pcre=100Pyg, (b) Pcre=5Pyg, and (c) Pcre=Pyg .......................................................................... 774H334
333HFigure 8.18 Comparison of simulated test strengths to predictions for columns with local buckling-controlled failures as a function of local slenderness (tested strength is normalized by Pne)............................................................................................................... 775H345
334HFigure 8.19 Comparison of simulated test strengths to predictions for columns with local buckling-controlled failures as a function of local slenderness (tested strength is normalized by Pyg) .............................................................................................................. 776H346
335HFigure 8.20 Comparison of simulated test strengths to predictions for columns with distortional buckling-controlled failures as a function of distortional slenderness ........................ 777H347
336HFigure 8.21 Comparison of simulated test strengths to predictions for columns with global buckling-controlled failures (i.e., no local interaction) as a function of global slenderness ........................................................................................................................... 778H348
337HFigure 8.22 Test-to-predicted ratios for local buckling-controlled simulated column failures as a function of local slenderness ............................................................................................. 779H349
338HFigure 8.23 Test-to-predicted ratios for distortional buckling-controlled simulated column failures as a function of distortional slenderness ........................................................... 780H350
339HFigure 8.24 Test-to-predicted ratios for simulated global buckling-controlled column failures (i.e., no local buckling interaction) as a function of global slenderness ............................... 781H351
340HFigure 8.25 Test-to-predicted ratios for simulated local buckling-controlled column failures as a function of net cross-sectional area to gross cross-sectional area ................................. 782H352
341HFigure 8.26 Test-to-predicted ratios for simulated distortional buckling-controlled column failures as a function of net cross-sectional area to gross cross-sectional area ........... 783H353
342HFigure 8.27 Test-to-predicted ratios for simulated global buckling-controlled column failures as a function of net cross-sectional area to gross cross-sectional area ................................. 784H354
343HFigure 8.28 Test-to-predicted ratios for simulated local buckling-controlled column failures as a function of column length, L, to flat web width, h ......................................................... 785H355
344HFigure 8.29 Test-to-predicted ratios for simulated distortional buckling-controlled column failures as a function of column length, L, to flat web width, h .................................... 786H356
345HFigure 8.30 Test-to-predicted ratios for simulated global buckling-controlled column failures as a function of column length, L, to web width, h ................................................................ 787H357
346HFigure 8.31 Comparison of experimental test strengths to predictions for columns with local buckling-controlled failures as a function of local slenderness (tested strength is normalized by Pne)............................................................................................................... 788H360
347HFigure 8.32 Comparison of experimental test strengths to predictions for columns with local buckling-controlled failures as a function of local slenderness (tested strength is normalized by Py) ................................................................................................................ 789H361
348HFigure 8.33 Comparison of experimental test strengths to predictions for columns with distortional buckling-controlled failures as a function of distortional slenderness .. 790H362
349HFigure 8.34 Comparison of experimental test strengths to predictions for columns with global buckling-controlled failures as a function of global slenderness ................................. 791H363
xx
350HFigure 8.35 Test-to-predicted ratios for experiment local buckling-controlled column failures as a function of local slenderness ............................................................................................. 792H364
351HFigure 8.36 Test-to-predicted ratios for experiment distortional buckling-controlled column failures as a function of distortional slenderness .......................................................... 793H365
352HFigure 8.37 Test-to-predicted ratios for experiment global buckling-controlled column failures as a function of global slenderness ........................................................................................ 794H366
353HFigure 8.38 Test-to-predicted ratios for experiment local buckling-controlled column failures as a function of net cross-sectional area to gross cross-sectional area ................................. 795H367
354HFigure 8.39 Test-to-predicted ratios for experiment distortional buckling-controlled column failures as a function of net cross-sectional area to gross cross-sectional area ........... 796H368
355HFigure 8.40 Test-to-predicted ratios for experiment global buckling-controlled column failures as a function of net cross-sectional area to gross cross-sectional area .............................. 797H369
356HFigure 8.41 Test-to-predicted ratios for experiment local buckling-controlled column failures as a function of column length, L, to flat web width, h ......................................................... 798H370
357HFigure 8.42 Test-to-predicted ratios for experiment distortional buckling-controlled column failures as a function of column length, L, to flat web width, h .................................... 799H371
358HFigure 8.43 Test-to-predicted ratios for experiment global buckling-controlled column failures as a function of column length, L, to web width, h ............................................................. 800H372
359HFigure 8.44 ABAQUS simulated beam experiments boundary conditions and application of loading .................................................................................................................................. 801H375
360HFigure 8.45 SSMA 800S162-43 beam with web holes considered in the DSM local buckling study ............................................................................................................................................... 802H378
361HFigure 8.46 Comparison of simulated beam strengths (Inet/Ig=1.0, no holes) to (a) the existing DSM local buckling design curve and to (b) the proposed DSM local buckling curve for beams with holes ................................................................................................................. 803H380
362HFigure 8.47 Comparison of simulated beam strengths (Inet/Ig=0.95) to (a) the existing DSM local buckling design curve and to (b) the proposed DSM local buckling curve for beams with holes ............................................................................................................................. 804H380
363HFigure 8.48 Comparison of simulated beam strengths (Inet/Ig=0.90) to (a) the existing DSM local buckling design curve and to (b) the proposed DSM local buckling curve for beams with holes ............................................................................................................................. 805H381
364HFigure 8.49 Comparison of simulated beam strengths (Inet/Ig=0.85) to (a) the existing DSM local buckling design curve and to (b) the proposed DSM local buckling curve for beams with holes ............................................................................................................................. 806H381
365HFigure 8.50 SSMA 550S162-54 structural stud failure mode transition from distortional buckling to yielding at the net section .............................................................................................. 807H383
366HFigure 8.51 Comparison of simulated beam strengths (Inet/Ig=1.0) to (a) the existing DSM distortional buckling design curve and to (b) the proposed DSM distortional buckling curve for beams with holes ............................................................................... 808H384
367HFigure 8.52 Comparison of simulated beam strengths (Inet/Ig=0.95) to (a) the existing DSM distortional buckling design curve and to (b) the proposed DSM distortional buckling curve for beams with holes ............................................................................... 809H384
368HFigure 8.53 Comparison of simulated beam strengths (Inet/Ig=0.90) to (a) the existing DSM distortional buckling design curve and to (b) the proposed DSM distortional buckling curve for beams with holes ............................................................................... 810H385
369HFigure 8.54 Comparison of simulated test strengths to predictions for laterally braced beams with local buckling-controlled failures as a function of local slenderness ........................... 811H394
370HFigure 8.55 Comparison of simulated test strengths to predictions for laterally braced beams with distortional buckling-controlled failures as a function of distortional slenderness .. 812H395
371HFigure 8.56 Test-to-predicted ratios for local buckling-controlled simulated laterally braced beam failures as a function of local slenderness ........................................................................ 813H396
xxi
372HFigure 8.57 Test-to-predicted ratios for distortional buckling-controlled simulated laterally braced beam failures as a function of distortional slenderness .................................... 814H397
373HFigure 8.58 Test-to-predicted ratios for simulated local buckling-controlled laterally braced beam failures as a function of net cross-sectional moment of inertia to gross cross-sectional moment of inertia ................................................................................................................ 815H398
374HFigure 8.59 Test-to-predicted ratios for simulated distortional buckling-controlled laterally braced beam failures as a function of net cross-sectional moment of inertia to gross cross-sectional moment of inertia ..................................................................................... 816H399
375HFigure 8.60 Test-to-predicted ratios for simulated local buckling-controlled laterally braced beam failures as a function of column length, L, to beam depth, H ....................................... 817H400
376HFigure 8.61 Test-to-predicted ratios for simulated distortional buckling-controlled laterally braced beam failures as a function of column length, L, to H ....................................... 818H401
377HFigure 8.62 Comparison of experimental test strengths to predictions for laterally braced beams with local buckling-controlled failures as a function of local slenderness .................. 819H404
378HFigure 8.63 Comparison of experimental test strengths to predictions for laterally braced beams with distortional buckling-controlled failures as a function of distortional slenderness ............................................................................................................................................... 820H405
379HFigure 8.64 Test-to-predicted ratios for experimental local buckling-controlled laterally braced beam failures as a function of net cross-sectional moment of inertia to gross cross-sectional moment of inertia ............................................................................................... 821H406
380HFigure 8.65 Test-to-predicted ratios for experimental distortional buckling-controlled laterally braced beam failures as a function of net cross-sectional moment of inertia to gross cross-sectional moment of inertia ..................................................................................... 822H407
381HFigure C.1 CUFSM finite strip modeling definition for an unstiffened element in compression 823H435 382HFigure C.2 The plate buckling coefficient kA for an unstiffened element in compression (the
multiple curves represent 0≤ψA≤1 with a step of 0.1, 11 curves total) ......................... 824H436 383HFigure C.3 The fitted curve for kA is a conservative predictor when Lhole/yA≤2 ............................... 825H437 384HFigure C.4 CUFSM finite strip modeling definition for an unstiffened element with compression
on the free edge, tension on the simply-supported edge .............................................. 826H438 385HFigure C.5 Variation in plate buckling coefficient kB for an unstiffened element with ψB ranging
from 0 to 10 .......................................................................................................................... 827H438 386HFigure C.6 Curve fit to minimum kB for ψB ranging from 0 to 10 ..................................................... 828H439 387HFigure C.7 Family of curves used to simulate boost in kB when Lhole/hB≤2, ψB ranges from 0 to 10
............................................................................................................................................... 829H440 388HFigure E.1 Long column with two holes spaced symmetrically about the longitudinal midline. 830H445
xxii
List of Tables 389HTable 3.1 Plate widths corresponding to SSMA structural stud designations ..................................................... 831H26 390HTable 3.2 Parameter ranges in stiffened element verification study. .................................................................... 832H36 391HTable 3.3 Parameter range for stiffened element verification study with offset holes. ...................................... 833H41 392HTable 3.4 Parameter ranges considered for stiffened elements in bending with holes. ...................................... 834H44 393HTable 3.5 Study parameter limits for stiffened element in bending (Y/h=0.50) with offset holes .................... 835H46 394HTable 3.6 Parameter range for study of regularly-spaced holes on unstiffened elements. ................................ 836H59 395HTable 3.7 Parameter range considered for unstiffened element study with offset holes ................................... 837H61 396HTable 4.1 SSMA structural stud and plate dimensions ........................................................................................... 838H70 397HTable 4.2 Summary of column experimental data .................................................................................................. 839H79 398HTable 4.3 Fixed-fixed column experiment dimensions and material properties ................................................. 840H82 399HTable 4.4 Fixed-fixed column experiment elastic buckling properties ................................................................. 841H83 400HTable 4.5 Weak-axis pinned column experiment dimensions and material properties ..................................... 842H84 401HTable 4.6 Weak-axis pinned column experiment elastic buckling properties ..................................................... 843H84 402HTable 4.7 Parameter ranges for fixed-fixed and weak-axis pinned column specimens with holes .................. 844H85 403HTable 4.8 DSM prequalification limits for C-sections ............................................................................................. 845H85 404HTable 4.9 DSM prequalification limits for beam C-sections ................................................................................. 846H143 405HTable 4.10 Parameter ranges for beam specimens with holes ............................................................................. 847H143 406HTable 4.11 Beam experiment cross-section dimensions, material properties, and tested strengths ............... 848H144 407HTable 4.12 Beam experiment elastic buckling results ........................................................................................... 849H145 408HTable 5.1 FSM local and distortional buckling half-wavelengths for nominal 362S162-33 and 600S162-33
cross-sections .................................................................................................................................................... 850H162 409HTable 5.2 Voltage conversion factors for column test instrumentation .............................................................. 851H166 410HTable 5.3 Summary of measured cross section dimensions ................................................................................. 852H175 411HTable 5.4 Summary of measured lip-flange and flange-web cross section angles ............................................ 853H176 412HTable 5.5 Specimen bare steel and zinc coating thicknesses ................................................................................ 854H178 413HTable 5.6 Measured column specimen length ....................................................................................................... 855H181 414HTable 5.7 Specimen end flatness .............................................................................................................................. 856H182 415HTable 5.8 Measured slotted hole dimensions and locations ................................................................................ 857H183 416HTable 5.9 Initial web imperfections (deviations from the average elevation of the web) ................................ 858H185 417HTable 5.10 Specimen gross centroid and offset from applied load during tests ............................................... 859H187 418HTable 5.11 Summary of out-of-straightness calculations ..................................................................................... 860H190 419HTable 5.12 Voltage conversion factors for tensile coupon testing ....................................................................... 861H194 420HTable 5.13 Summary of column specimen steel yield stress ................................................................................ 862H195 421HTable 5.14 Column specimen steel yield stress statistics ...................................................................................... 863H196 422HTable 5.15 Critical elastic buckling loads, influence of holes on elastic buckling ............................................. 864H200 423HTable 5.16 Specimen ultimate strength results ...................................................................................................... 865H206 424HTable 6.1 Statistics of the residual stresses in roll-formed members .................................................................. 866H249 425HTable 6.2 Radial location in the coil that minimizes the sum of the mean square prediction error for roll-
formed members .............................................................................................................................................. 867H252 426HTable 7.1 Summary of nonlinear finite element models and associated solution controls .............................. 868H271 427HTable 7.2 Local and distortional imperfection magnitudes ................................................................................. 869H297 428HTable 7.3 Out-of-straightness imperfection magnitudes ...................................................................................... 870H298 429HTable 7.4 Comparison of nonlinear FE simulation peak loads to experiments ................................................. 871H303 430HTable 8.1 DSM test-to-predicted statistics for column simulations .................................................................... 872H344 431HTable 8.2 DSM test-to-predicted ratio statistics for column experiments .......................................................... 873H359 432HTable 8.3 DSM test-to-predicted ratio statistics for column experiments (stub columns only) ....................... 874H359 433HTable 8.4 DSM test-to-predicted statistics for laterally braced beam simulations ............................................ 875H393 434HTable 8.5 DSM test-to-predicted ratio statistics for column experiments .......................................................... 876H403
1
Chapter 1 0BIntroduction
The goal of this research work is to develop a general design method for cold‐
formed steel structural members with holes. Cold‐formed steel beams and columns are
typically manufactured with perforations. For example, in low and midrise
construction, holes are prepunched in structural studs to accommodate the passage of
utilities in the walls and ceilings of buildings as shown in 877HFigure 1.1. In cold‐formed
steel storage rack columns, perforation patterns are provided to allow for variable shelf
configurations as shown in 878HFigure 1.2. (Members with discrete holes, for example C‐
sections with punched holes as shown in 879HFigure 1.1, are the research focus in this thesis,
although many of the tools and methods developed here can be extended to perforation
patterns in storage racks with additional research effort.) Existing design procedures for
cold‐formed steel members with holes are limited to certain hole sizes, shapes, and
configurations. These limitations can hamper an engineer’s design flexibility and
decrease the reliability of cold‐formed steel components where holes exceed these
prescriptive limits.
2
Figure 1.1 Perforations are provided in structural studs to accommodate utilities in the walls of buildings
Figure 1.2 Hole patterns in storage rack columns The basic framework of the design procedure developed in this thesis is the Direct
Strength Method (DSM) (AISI‐S100 2007, Appendix 1). DSM is relatively new and
represents a major advancement in cold‐formed steel design because it provides
engineers and cold‐formed steel manufacturers with the tools to predict the strength of a
member with any general cross‐section. Cold‐formed steel members are manufactured
from thin sheet steel, and therefore member resistance is influenced by cross‐section
instabilities (e.g., plate buckling and distortion of open cross‐sections) in addition to the
global buckling influence considered in thicker hot‐rolled steel sections. DSM explicitly
defines the relationship between elastic buckling and load‐deformation response with
empirical equations to predict ultimate strength.
3
To calculate the capacity of a cold‐formed steel member with DSM, the elastic
buckling properties of a general cold‐formed steel cross‐section are obtained from an
elastic buckling curve. The curve can be generated with software employing the finite
strip method to perform a series of eigenbuckling analyses over a range of buckled half‐
wavelengths. In this research work the freely available program CUFSM is utilized
(Schafer and Ádàny 2006). An example of an elastic buckling curve is provided in
880HFigure 1.3 for a cold‐formed steel C‐section column and highlights the three categories of
elastic buckling considered in DSM – local buckling, distortional buckling, and global
buckling. Local buckling occurs as plate buckling of individual slender elements in a
cross‐section. Distortional buckling exists only for open cross‐sections such as a C‐
section, where the compressed flanges buckle inward or outward along the length of a
member. Global buckling, also known as Euler buckling, defines buckling of the full
member at long half‐wavelengths including both flexural and flexural‐torsional effects
(and lateral‐torsional effects in beams).
The critical elastic buckling loads associated with local, distortional, and global
buckling – Pcrl, Pcrd, and Pcre for columns (Mcrl, Mcrd, and Mcre for beams), can be obtained
directly from the elastic buckling curve. The critical elastic buckling loads are then used
to predict the ultimate strength with three empirical design curves presented in 881HFigure
1.4 to 882HFigure 1.6 for cold‐formed steel columns. (The current DSM column design
equations for members without holes are also provided in these figures.) The local,
distortional, and global slenderness of a member (λl, λd, λc) are calculated from the
critical elastic buckling loads, defining a member’s sensitivity to each type of buckling at
4
failure (high slenderness corresponds to high sensitivity, low slenderness to low
sensitivity). The nominal resistances (Pnl, Pnd, and Pne) are obtained by inserting the
slenderness magnitudes into the DSM design equations. The minimum of the local,
distortional, and global nominal strengths is taken as the strength of the member.
100 101 102 1030
2
4
6
8
10
12
14
16
18
20
half-wavelength
load
fact
or
half-wavelength (in.)
Pcr
(kip
s)
Local buckling Distortional buckling
Global buckling
lcrP creP
crdP
Pcrl
Pcrd
Pcre
Figure 1.3 Column elastic buckling curve generated with CUFSM
0 0.5 1 1.5 2 2.5 3 3.5 40
0.5
1
1.5
Global slenderness, λc=(Py/Pcre)0.5
Pne
/Py
Flexural, Torsional, or Torsional-Flexural Buckling The nominal axial strength, Pne, for flexural or torsional- flexural buckling is
for 5.1c ≤λ Pne = yP658.02c ⎟
⎠⎞⎜
⎝⎛ λ
for λc > 1.5 crey2c
ne P877.0P877.0P =⎟⎟⎠
⎞⎜⎜⎝
⎛
λ=
where λc = crey PP
Py = AgFy Pcre= Critical elastic global column buckling load Ag = gross area of the column
Figure 1.4 DSM global buckling failure design curve and equations
5
0 0.5 1 1.5 2 2.5 3 3.5 40
0.5
1
1.5
local slenderness, λl=(Pne/Pcrl)
0.5
Pn l
/Pne
Local Buckling The nominal axial strength, Pnl, for local buckling is
for λl 776.0≤ Pnl = Pne
for λl > 0.776 Pnl = ne
4.0
ne
cr4.0
ne
cr PPP
PP
15.01 ⎟⎟⎠
⎞⎜⎜⎝
⎛⎥⎥
⎦
⎤
⎢⎢
⎣
⎡⎟⎟⎠
⎞⎜⎜⎝
⎛− ll
where λl = lcrne PP Pcrl = Critical elastic local column buckling load Pne = Nominal axial strength for global buckling
Local buckling interacts with global buckling at failure
Global failure
Figure 1.5 DSM local buckling failure design curve and equations
0 0.5 1 1.5 2 2.5 3 3.5 40
0.5
1
1.5
Distortional slenderness, λd=(Py/Pcrd)0.5
Pnd
/Py
Distortional Buckling The nominal axial strength, Pnd, for distortional buckling is
for λd 561.0≤ Pnd = Py
for λd > 0.561 Pnd = y
6.0
y
crd
6.0
y
crd PPP
PP
25.01 ⎟⎟⎠
⎞⎜⎜⎝
⎛
⎟⎟⎟
⎠
⎞
⎜⎜⎜
⎝
⎛
⎟⎟⎠
⎞⎜⎜⎝
⎛−
where λd = crdy PP
Pcrd = Critical elastic distortional column buckling load Py = Column yield strength
Figure 1.6 DSM distortional buckling failure design curve and equations
6
This research aims to extend the appealing generality of DSM to cold‐formed steel
members with perforations. The primary research goals are to study and quantify the
influence of holes on the elastic buckling of cold‐formed steel beams and columns and
then to develop modifications to the existing DSM design equations which relate elastic
buckling to ultimate strength. The research plan is implemented in three phases:
Phase I (Chapters 2‐4)
1. Study the influence of holes on the elastic buckling of thin plates, and then on cold‐formed steel beams and columns.
2. Evaluate the viability of DSM for members with holes by comparing existing
experiments on members with holes to the current DSM specification. Phase II ( Chapters 5‐7)
1. Conduct experiments on cold‐formed steel columns to observe the influence of holes on ultimate strength and post‐buckling response.
2. Define and validate a nonlinear finite element modeling protocol through parameter
studies on thin plates and comparison to experimental results. Phase III (Chapters 7‐8)
1. Formalize the relationship between elastic buckling and ultimate strength for members with holes using nonlinear finite element simulations and existing data.
2. Modify the current DSM specification to account for members with holes Phase I research is primarily focused on elastic buckling. Chapter 2 describes
preliminary thin shell finite element eigenbuckling studies which are used to evaluate
the accuracy of different shell element types in ABAQUS and to define finite element
meshing guidelines. Chapter 3 extends this elastic buckling research with eigenbuckling
analyses of typical cross‐sectional elements considered in cold‐formed steel design. For
7
example, a stiffened element is a simply‐supported plate used to model the web of a
cold‐formed steel C‐section and an unstiffened element is a plate simply‐supported on
three edges and free on the fourth edge to simulate the behavior of the free leg of a cold‐
formed steel hat section. Chapter 4 examines the elastic buckling of full cold‐formed
steel beams and columns with holes and develops useful simplified methods to predict
elastic buckling, including the influence of holes, without finite element analysis. The
elastic buckling properties of existing beam and column experiments are also calculated
and merged with the tested strengths into a database. This database is employed near
the end of the project to validate the proposed modifications to the DSM design
equations for members with holes.
Phase II marks a shift from elastic buckling to the study of the influence of holes on
load‐deformation response and ultimate strength. Chapter 5 describes an experimental
program on short and intermediate length cold‐formed steel columns with holes.
Chapter 6 initiates the development of a nonlinear finite element protocol with a
significant effort to define the residual stresses and initial plastic strains from the
manufacturing process. The capabilities of the commercial program ABAQUS
(ABAQUS 2007a) are explored at the beginning of Chapter 7 with preliminary nonlinear
finite element simulation studies on rectangular plates with holes. The experimental
results from Chapter 5 are then employed in Phase III to fully develop and verify the
modeling protocol. The research culminates in Chapter 8 with the development of a
database of simulated tests which are used in combination with existing experimental
data to validate the DSM design method for cold‐formed steel members with holes.
8
Chapter 2 1BThin-shell finite element modeling in ABAQUS
A set of ABAQUS modeling guidelines is formalized in this chapter to provide a
consistent methodology for the finite element studies conducted in this thesis research.
Finite element eigenbuckling analysis is a valuable tool for studying the elastic buckling
properties of thin‐walled structures. The accuracy of an analysis is influenced by
decisions made while implementing the finite element model, including the choice of
finite element type and the meshing geometry and density. Studies are presented here
which compare finite element eigenbuckling predictions of plate buckling problems to
known theoretical solutions. The eigenbuckling analyses are performed with the
commercial finite element program ABAQUS (ABAQUS 2007a). The accuracy of
ABAQUS thin shell elements are evaluated, and finite element convergence studies are
presented to identify limits on element aspect ratio. Rules for modeling rounded corners
and meshing around holes are also provided with supporting elastic buckling studies.
9
2.1 22BComparison of ABAQUS thin‐shell elements
Three ABAQUS finite elements commonly employed in the elastic buckling analysis
of thin‐walled structures are the S9R5, S4, and S4R elements as shown in 883HFigure 2.1. The
S4 and the S4R finite elements are four node general purpose shell elements valid for
both thick and thin shell problems (ABAQUS 2007a). Both elements employ linear
shape functions to interpolate deformation between nodes. The S9R5 element is a
doubly‐curved thin shell element with nine nodes derived with shear flexible Mindlin
strain definitions and Kirchoff constraints (classical plate theory with no transverse
shear deformation) enforced as penalty functions (Schafer 1997). This element employs
quadratic shape functions (resulting from the increase in the number of nodes from 4 to
9) which provide two important benefits when modeling thin‐walled structures: (1) the
ability to define initially curved geometries and (2) the ability to approximate a half sine
wave with just one element. The “5” in S9R5 denotes that each element node has 5
degrees of freedom (three translational, two rotational) instead of 6 (three translational,
three rotational). The rotation of a node about the axis normal to the element mid‐
surface is removed from the element formulation to improve computational efficiency.
The “R” in the S9R5 (and S4R) designation denotes that the calculation of the element
stiffness is not exact; the number of Gaussian integration points is reduced to improve
computational efficiency and to avoid shear locking. This “reduced integration”
approach underestimates element stiffness and sometimes results in artificial element
deformation modes with zero strain across the element, commonly referred to as
“hourglass” modes (Schafer 1997). The accuracy of eigenbuckling finite element models
10
are compared here for each of these ABAQUS element types against the exact solutions
for two common plate buckling problems.
S4/S4R S9R5
Figure 2.1 ABAQUS S4\S4R shell element with four nodes and a linear shape function, ABAQUS S9R5 shell element with nine nodes and a quadratic shape function
1.1 52BModeling accuracy for a stiffened element Elastic buckling analyses of a stiffened element were performed in ABAQUS to
compare the accuracy of the ABAQUS S9R5, S4, and S4R elements against a known
solution. A stiffened element is a common term used in thin‐walled structures to
describe a cross‐sectional element restrained on both edges (see 884HFigure 3.1) which is
approximated as a thin simply‐supported plate (with sides free to wave) and loaded
uniaxially as shown in 885HFigure 2.2.
L
h
fcr
Figure 2.2 Buckled shape of a stiffened plate
11
The theoretical buckling stress for a stiffened element is:
( )2
2
2
cr ht
112E
kf ⎟⎠⎞
⎜⎝⎛
−=
νπ
, (2.1)
where h is the width of the plate, E is the modulus of elasticity of the plate material, ν is
the Poisson’s ratio, and t is the thickness of the plate.
The buckling coefficient k is:
22
⎟⎟⎠
⎞⎜⎜⎝
⎛+=
mhLn
Lmhk , (2.2)
where L is the length of the plate and m and n are the number of half‐wavelengths in the
longitudinal and transverse directions, respectively (Chajes 1974). In 886HFigure 2.2, m=4
and n=1.
Plate buckling coefficients (k) are approximated in ABAQUS by performing
eigenbuckling analyses of stiffened elements with ABAQUS S4, S4R, and S9R5 elements.
The element aspect ratio is set at 8:1 for the S9R5 element and 4:1 for the S4R and S4
elements to ensure a consistent comparison between finite element models (i.e., similar
numbers of nodes and computational demand). These particular element aspect ratios
were also chosen because they are expected to be towards the upper limit of what is
required to discretize the geometry of cold‐formed steel members (especially at rounded
corners where the element aspect ratio can be quite high). The plate thickness is set to
t=0.0346 in. E=29500 ksi and ν=0.30 for all finite element models. The ABAQUS
boundary and loading conditions are implemented as shown in 887HFigure 3.1.
12
888HFigure 2.3 compares the theoretical k from Eq. 889H(2.1) to the ABAQUS buckling
coefficients for varying plate aspect ratios (L/h). The S9R5 element performs accurately
over the range of element aspect ratios considered, with a maximum error of 1.3 percent.
The S4 and S4R elements are not as accurate, with maximum errors of 11.4 percent and
9.7 percent, respectively. The accuracy of the plate models with S4 and S4R elements
increase with increasing plate aspect ratio, which indirectly implies that solution
accuracy increases as the number of elements per half‐wave increase (in the loaded
direction). This hypothesis is consistent with the element formulations, since the S9R5
element uses a quadratic shape function to estimate displacements (and can therefore
capture the half‐sine wave of a buckled plate with as little as one element) and the S4
and S4R elements use linear shape functions (requiring at least three elements to
coarsely estimate the shape of a half sine wave). The S4R element is observed to be
slightly less stiff than the S4 element in 890HFigure 2.3, which is hypothesized to occur as a
result of the reduced integration stiffness approximation.
Comparing the number of elements required to model a buckled half‐wave is a
more useful indicator of mesh density and model accuracy than just the element aspect
ratio alone. 891HFigure 2.4 verifies this supposition by demonstrating the improvement in
modeling accuracy for a stiffened element as the number of finite elements per square
half‐wave increase. The S4 element experiences membrane locking when the number of
elements per half wave is less than 2, resulting in exceedingly unconservative values for
k. The S4R avoids this membrane locking with a reduced integration scheme that
assumes the membrane stiffness is constant in the element (ABAQUS 2007a).
13
Regardless, the accuracy of the S4R element degrades when less than 5 elements per
half‐wavelength are used and neither four node element (i.e., the S4 or the S4R) is able to
capture the sinusoidal shape of the buckled half‐wave with less than three elements per
buckled half‐wave. The S9R5 accurately predicts the shape of the buckled half‐wave
and the buckling coefficient k with just one element. k is within 2.1 percent of the
theoretical value for one element per half‐wave and reduces to 0.1 percent for two
elements per half‐wave.
0 0.5 1 1.5 2 2.5 3 3.5 43.5
4
4.5
5
5.5
6
plate aspect ratio, L/h
plat
e bu
cklin
g co
effic
ient
, k
S4S4RS9R5Theory
S9R5 (8:1) S4\S4R (4:1)
L/h=2 shown
Figure 2.3 Accuracy of ABAQUS S9R5, S4, and S4R elements for a stiffened element with varying aspect ratios, 8:1 finite element aspect ratio for the S9R5 element, 4:1 element aspect ratio for the S4 and S4R
elements
14
0 2 4 6 8 10 12 14 16 18 203
3.5
4
4.5
5
5.5
6
6.5
7
number of elements per buckled half-wave
buck
ling
coef
f., k
S4S4RS9R5
Figure 2.4 Accuracy of S4, S4R, and S9R5 elements as a function of the number of elements provided per buckled half‐wavelength, stiffened element, square waves (k=4)
1.2 53BModeling accuracy for an unstiffened element An unstiffened element is another common cross‐section component considered in
the elastic buckling of thin‐walled cross‐sections (see 892HFigure 3.1), the behavior of which
is conservatively approximated as a plate simply‐supported on three sides and free on
the fourth side parallel to the direction of a uniaxially applied stress. The buckled shape
of an unstiffened element is depicted in 893HFigure 2.5.
L
h
fcr
15
Figure 2.5 Buckled shape of an unstiffened element, m=1 shown
The theoretical buckling coefficient k of an unstiffened element can be calculated with
the numerical solution of the following equations (Timoshenko 1961):
( ) ( )hL
mhL
m βπνβααπναβ tanhtanh 2
222
2
222
⎟⎟⎠
⎞⎜⎜⎝
⎛+=⎟⎟
⎠
⎞⎜⎜⎝
⎛− , (2.3)
2
1
212
2
22
⎟⎟⎠
⎞⎜⎜⎝
⎛+= k
Lhm
Lm ππα , and
21
212
2
22
⎟⎟⎠
⎞⎜⎜⎝
⎛+−= k
Lhm
Lm ππβ . (2.4)
894HFigure 2.6 compares the theoretical to predicted k versus the number of S9R5
elements provided along the length L of an unstiffened element. The plate dimensions
are held constant at L/h= 4, while the element aspect ratio is varied from 1:1 to 64:1. The
S9R5 element produces an error of 4.3 percent with an element aspect ratio of 16:1 and
an error of 1.0 percent with an element aspect ratio of 8:1.
0 10 20 30 40 50 60 700
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
number of S9R5 elements along length of plate
buck
ling
coef
ficie
nt, k
64:1
32:1
16:1
8:1 2:1 1:14:1
element aspect ratio (typ.)
k=0.486 when L/h=4
Figure 2.6 Accuracy of S9R5 elements as the number of finite elements provided along an unstiffened
element varies, L/h=4
16
2.2 23BModeling holes in ABAQUS The ability to incorporate holes into the geometry of an ABAQUS finite element
model is a key prerequisite to studying the influence of holes on the structural behavior
of cold‐formed steel structural members. To clear this hurdle, custom Matlab code was
written by the author which generates a finite element mesh of a plate containing a hole
(Mathworks 2007). The code discretizes the geometry around a hole by creating layers
of S9R5 elements as shown in 895HFigure 2.7 for a slotted hole, a circular hole, and a square
hole. (See 896HAppendix A for a description of the custom mesh generation program.
Additional Matlab tools were developed to integrate the hole mesh geometry into an
existing finite element model.) The discretization results in S9R5 elements with
opposite edges which are not initially parallel. The initial geometry of 9 node
quadrilateral elements without parallel edges can be defined without loss of accuracy as
long as the midline nodes remain centered between the corner nodes (Cook 1989), which
is an advantage over the S4 and S4R elements. ABAQUS recommends that the angle
between isoparametric lines (i.e., corner angles of an element) should not be less than 45
degrees or greater than 135 degrees to ensure accurate numerical integration of the
element stiffness matrix (ABAQUS 2007a). This limit coincides with the minimum and
maximum S9R5 corner angles for the elements at the bisection of the 90 degree plate
corners as shown in 897HFigure 2.8.
This study establishes ABAQUS S9R5 finite element mesh guidelines for plates with
holes by studying the convergence of the elastic stability solution as element aspect ratio
17
varies. 898HFigure 2.7 provides the typical mesh layout and summarizes the plate
dimensions considered in this study. The plate is modeled as a stiffened element in
ABAQUS, simply supported on four sides and loaded uniaxially in compression (see
899HFigure 3.5 for the ABAQUS implementation of the boundary and loading conditions).
The plate thickness t=0.0346 in., E=29500 ksi, and ν=0.30 for all finite element models.
L=6.0 in. L=3.4 in.
h=3.4 in.hhole=1.5 in.
Lhole=4.0 in. Figure 2.7. Finite element mesh and plate dimensions: slotted, rectangular, and circular holes
The convergence of the elastic buckling solution for the plates with holes is studied
by varying the S9R5 element aspect ratio (a:b) at the bisection of the plate corners as
shown in 900HFigure 2.8, where a and b are defined as
2layers
hole
Nhha −
= , elem
hole
Nhb = . (2.5)
The aspect ratio is varied by increasing the number of finite element layers around the
hole (Nlayers) while the maintaining the number of edge elements (Nelem) constant (i.e., the
mesh density increases but the element corner angles remain constant).
45°
135°
a
b
Element aspect ratio is a:bNlayers number of element layers around hole
Nelem number of edge elements
Nlayers =4
Nelem =10
18
Figure 2.8 Hole discretization using S9R5 elements
901HFigure 2.9 demonstrates that the critical elastic buckling stress for the lowest
buckling mode (a half sine wave in this case) for all hole types converges to a constant
magnitude when a:b is between 0.5 and 2. This result is employed as a modeling
guideline for the research work in this thesis with the expressions for a and b in Eq. 902H(2.5):
212
5.0 ≤⎟⎟⎠
⎞⎜⎜⎝
⎛−≤
holelayers
elem
hh
NN
. (2.6)
0 0.5 1 1.5 2 2.5 3 3.5 40
0.2
0.4
0.6
0.8
1
1.2
S9R5 element aspect ratio, a/b
f cr,h
ole/f cr
,no
hole
circular holesquare holeSSMA slotted hole
Figure 2.9 The critical elastic buckling stress converges to a constant magnitude when the S9R5 element aspect ratio a/b is between 0.5 and 2 and element corner angles are skewed
2.3 24BModeling Rounded Corners in ABAQUS The S9R5 element can be defined with an initial curved geometry in ABAQUS which
makes it convenient for modeling rounded corners of a cold‐formed steel cross‐section.
ABAQUS recommends that the initial element curvature should be less than 10 degrees,
19
where curvature of an S9R5 element is defined as the angle subtended by the nodal
normal and the average element normal as shown in 903HFigure 2.10. The derivation in
904HFigure 2.10 demonstrates that this curvature recommendation is met when five or more
S9R5 elements form the 90 degree corner. This limit is unfavorable from a modeling
perspective because the element aspect ratio increases as the number of elements around
the corner increase, another potential source of accuracy degradation. Also, for a finite
element model with four 90 degree corners (e.g., a cold‐formed steel lipped C‐section),
increasing the number of elements at a corner can result in a considerable increase in
computational demand if the corner elements comprise a large proportion of the total
number of elements in a cross‐section.
Node normal (typ.)
Average S9R5 element normal
ABAQUS recommends α ≤10 degrees to limit initial S9R5 element curvature
α
r
S
5.4,2
2,
18
18,
90
90
≥∴≥
=≤
≤=
elemelem NS
SN
rSrS
rS
ππ
πααS
Figure 2.10 ABAQUS S9R5 initial curvature limit requires at least five elements to model corner A parameter study was conducted to evaluate the influence of the number of S9R5
elements making up a 90 degree corner on the critical elastic buckling loads for local
buckling (Pcrl), distortional buckling (Pcrd), and global buckling (Pcre) of an SSMA 600S162‐
68 C‐section column. The number of corner elements were varied from 1 to 5, with the
associated S9R5 aspect ratio a:b varying from 5 to 22. The column length was held
constant at L=48 inches for all models to accommodate multiple local and distortional
half‐waves. The columns were loaded uniaxially and modeled with warping‐free ends
20
(CUFSM‐style boundary conditions) as shown in 905HFigure 4.2. E=29500 ksi and ν=0.30 for
all finite element models. Py is the squash load of the column calculated with the steel
yield stress Fy=50 ksi.
906HFigure 2.11 provides the typical mesh geometry of the column and compares a C‐
section corner modeled as a smooth surface with one S9R5 element and with three S9R5
elements. 907HFigure 2.12 demonstrates that the number of S9R5 corner elements has a
minimal influence on the elastic buckling behavior of the column, with a slight
decreasing trend (less than 1%) in critical elastic buckling load with increasing element
quantity. Mesh refinement at the corners does not influence solution accuracy because
elastic buckling deformation occurs primarily within the more flexible cross‐sectional
elements. If the simulation of sharp folding of the corners is required, such as in the case
of nonlinear finite element modeling to collapse, additional corner elements may be
warranted to accurately capture localized deformation gradients.
Figure 2.11 SSMA 600S162‐68 C‐section corner modeled with a) one S9R5 element, b) three S9R5 elements
21
0 1 2 3 4 5 60
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1
Number of corner S9R5 elements
Pcr
/Py
Global - flexural torsionalGlobal - weak axis flexureDistortionalLocal
Figure 2.12 The number of S9R5 corner elements has a minimal influence on the critical elastic buckling loads of an SSMA 600162‐68 C‐section column with L=48 in.
2.4 25BSummary of modeling guidelines
The S9R5 element will be implemented in this research work based on its versatility
and demonstrated accuracy. The results of the ABAQUS studies in this chapter form the
basis of the ABAQUS modeling guidelines below which will be implemented for both
eigenbuckling and nonlinear finite element studies in this thesis:
• A minimum of two S9R5 elements per half‐wavelength shall be provided in stiffened elements in the direction normal to the applied load (e.g., flanges and web of a lipped C‐section)
• The S9R5 element aspect ratio shall be less than or equal to 8:1 in unstiffened elements (e.g., flange lip in a C‐section)
• The S9R5 element aspect ratio shall be between 0.5 and 2.0 when modeling holes with the discretization scheme described in Section 908H2.2 (where the element sides are not perpendicular)
• For both stiffened and unstiffened elements, at least two S9R5 elements shall be provided in the direction perpendicular to the application of load
• Rounded corners shall be modeled with at least two S9R5 elements, and the element aspect ratio of these elements shall be less than or equal to 16:1.
22
Chapter 3 2BElastic buckling of cold-formed steel cross-sectional elements with holes
A simplified method for determining the elastic buckling properties of a thin‐walled
cross‐section is to evaluate the contribution of each element in the cross‐section
separately. This element‐by‐element evaluation is the basis of the effective width design
method for cold‐formed steel beams and columns and can also be employed as a
conservative predictor of the local critical elastic buckling load (Pcrl) when designing
cold‐formed steel members with the Direct Strength Method (AISI‐S100 2007, Appendix
1). The two common cross‐section element types in an open thin‐walled cross section
are stiffened and unstiffened elements, examples of which are provided in 909HFigure 3.1.
The boundary conditions of a stiffened element are conservatively approximated as a
simply‐supported plate. The unstiffened element is treated as a plate simply‐supported
on three sides and free on the fourth edge parallel to the application of load.
23
Web, stiffened element
Flange, stiffened element
Lip, unstiffenedelement
Figure 3.1 Stiffened and unstiffened elements in a lipped C‐section
The influence of holes on the elastic buckling behavior of stiffened and unstiffened
elements is evaluated in this chapter using thin shell finite element eigenbuckling
analysis. The presence of holes can modify the buckled mode shape of an element and
either increase or decrease its critical elastic buckling stress. Hole spacing and hole size
relative to element size are studied for both stiffened and unstiffened elements, and
approximate methods for predicting element critical elastic buckling stress are
developed and presented for use in design. The research results presented here will be
used as a framework for the elastic buckling studies of full cold‐formed steel structural
members with holes in 910HChapter 4.
3.1 26BPlate and hole dimensions
Plate and hole dimension nomenclature used throughout this chapter is summarized
in 911HFigure 3.2. The strips of plate between a hole and the plate edges will be referred to as
unstiffened strip “A” and unstiffened strip “B”, where the widths of these unstiffened
strips are hA and hB respectively as shown in 912HFigure 3.3. For stiffened elements in
bending, the neutral axis location is defined as Y in 913HFigure 3.4 and is measured from the
compressed edge of the plate.
24
C Hole
Detail AS
Lholehhole
Detail A
h
L+δhole
Plate with holes
Slotted hole
rhole
S/2
LC Plate
Figure 3.2 Element and hole dimension definitions
hA
hB
Unstiffened strip “A”
Unstiffened strip “B”
Detail A
Figure 3.3 Definition of unstiffened strip “A” and “B” for a plate with holes.
Tension edge
Detail A
Compressed edge
Y
Neutral axis
Figure 3.4 Definition of neutral axis location for stiffened elements in bending.
25
3.2 27BFinite element modeling assumptions
The elastic buckling behavior of stiffened and unstiffened elements with holes are
obtained with eigenbuckling analyses of plates in ABAQUS (ABAQUS 2007a). All
members are modeled with ABAQUS S9R5 reduced integration nine‐node thin shell
elements. The typical finite element aspect ratio is 1:1 and the maximum aspect ratio is
limited to 8:1 (refer to 914HChapter 2 for a discussion on ABAQUS thin shell finite element
types and finite element aspect ratio limits). Element meshing was performed with a
Matlab (Mathworks 2007) program written by the author (refer to 915HAppendix A for a
description of the program). The plate models are loaded from each end with stress
distributions applied as consistent nodal loads in ABAQUS. Converting a stress
distribution to consistent nodal loads for the S9R5 element requires a different
procedure than that followed for a 4‐node finite element (Schafer 1997). Cold‐formed
steel material properties are assumed as E=29500 ksi and ν=0.3 in all finite element
models.
3.3 28BStiffened element in uniaxial compression
3.1 54BBoundary and loading conditions
The stiffened element is modeled with simply‐supported boundary conditions and
loaded uniaxially with a uniform compressive stress as shown in 916HFigure 3.5.
26
perimeter supported in 2 (v = 0)
1
2
3transverse midline supported in 3 (w = 0)
longitudinal midline supported in 1 (u = 0)
Figure 3.5 ABAQUS boundary conditions and loading conditions for a stiffened element in uniaxial compression
3.2 55BInfluence of a single slotted hole
This study explores the influence of a single slotted hole on the elastic buckling
stress of a stiffened element. The plate length L is varied from three to twenty‐four
times the slotted hole length, Lhole, and the width of the plates are chosen to equal the flat
web widths of four common Steel Stud Manufacturers Association (SSMA) structural
studs listed in 917HTable 3.1 (SSMA 2001). The slotted hole has dimensions of hhole=1.5 in.,
Lhole=4 in., and rhole=0.75 in. Holes are always centered transversely between the unloaded
edges of the plate in this study. The plate thickness, t, is 0.0346 in.
Table 3.1 Plate widths corresponding to SSMA structural stud designations
SSMA hDesignation (in)250S162-33 2.28 0.66362S162-33 3.40 0.44600S162-33 5.78 0.26800S162-33 7.78 0.19
hhole/h
The results of this study are presented in 918HFigure 3.6 and demonstrate that as the
length of a stiffened element increases relative to the length of the hole, the critical
elastic buckling stress, fcr, converges to a constant magnitude which is either equal to or
lower than the buckling stress of a plate without a hole. The convergence occurs as L/Lhole
27
exceeds 5, suggesting that the influence of the hole is independent of the plate end
conditions beyond this length.
0 5 10 15 20 250
0.2
0.4
0.6
0.8
1
1.2
L/Lhole
f cr,h
ole/f cr
,no
hole
hhole/h=0.66
hhole/h=0.44
hhole/h=0.19
hhole/h=0.26
Figure 3.6 Influence of a slotted hole on the elastic buckling stress of a simply supported rectangular plate with varying length
When the hole is wide relative to the width of the plate (hhole/h=0.66) and L/Lhole is
small (see 919HFigure 3.6), the elastic buckling stress of the plate with the hole is as much as 7
percent higher than for a plate without a hole. This increase in stress is explained by the
buckled mode shapes in 920HFigure 3.7. The plate with the hole in 921HFigure 3.7a has a higher
elastic buckling stress than the plate without the hole in 922HFigure 3.7b because the natural
pattern of buckled waves is modified by the hole. The buckled cells adjacent to the hole
are shorter and therefore stiffer. The thin strips at the hole dampen buckling in this case
because they have an axial stiffness higher than the buckled cells away from the hole.
28
(a) (b)
Figure 3.7 Comparison of buckled shape and displacement contours for a rectangular plate with hhole/h=0.66 and L/Lhole =3, (a) with slotted hole and (b) without hole. Notice the change in length and quantity of buckled
cells with the addition of a slotted hole.
As the plate length increases past L/Lhole=5 for the smallest plate width
(hhole/h=0.66), the buckling stress converges to that of a plate without a hole. 923HFigure 3.8
demonstrates that for these long, slender stiffened elements the slotted hole dampens
buckling near the hole but does not appreciably change the natural half‐wavelength of
the buckled cells as was observed for the shorter plates in 924HFigure 3.7.
(a) (b)
Figure 3.8 Buckled shape of a simply supported plate (a) with a slotted hole and (b) without a hole. L=15Lhole , hhole/h=0.66. The slotted hole dampens buckling but does not significantly change the natural half‐
wavelength of the plate.
For plates with hhole/h less than 0.66, the slotted hole causes a decrease in the
elastic buckling stress which converges to a constant magnitude as the plate length
exceeds L/Lhole=5. 925HFigure 3.9a demonstrates that local buckling near the hole controls the
29
elastic buckling stress of these wider plates. The deformation at the hole results from the
localized reduction in transverse plate bending stiffness.
As plate length decreases below L/Lhole<5 and hhole/h=0.19, the influence of the hole
on the critical elastic buckling stress fluctuates as shown in 926HFigure 3.6. When the lowest
elastic buckling mode shape results in an odd number of half‐waves, the hole falls
within the central half‐wave and the critical elastic buckling stress decreases. For an
even number of half‐waves, the hole is located at the transition between two half sine‐
waves (because the hole is centered at the midlength of the plate), forcing the buckled
cells to shorten and increasing the critical elastic buckling stress.
(a) (b)
Figure 3.9 (a) Slotted hole causes local buckling (hhole/h=0.26), compared to (b) buckled cells at the natural half‐wavelength of the plate
3.3 56BInfluence of slotted hole spacing
The previous study demonstrated that the elastic buckling behavior of a stiffened
element with a single hole is sensitive to the size of the hole relative to the size of the
plate. The focus now shifts to the influence of multiple slotted holes on the elastic
buckling stress of a long fixed length stiffened element. In this study, slotted holes are
added one by one to a stiffened element (where L=24 Lhole) such that the center‐to‐center
spacing S varies as shown in 927HFigure 3.10.
30
SS/2 Lhole hholeh
Figure 3.10 Definition of center‐to‐center dimension for the slotted holes
As hole spacing decreases, the elastic buckling stress in 928HFigure 3.11 either increases
or decreases depending on the ratio of hole width to plate width. When there are many
large holes (hhole/h=0.66, S/ Lhole < 4), buckling is dampened at the holes and the buckled
cells shorten their lengths to form between adjacent holes (see 929HFigure 3.12 for buckled
shape). The decrease in buckled half‐wavelength causes an increase in elastic buckling
stress of the plate.
0 5 10 15 20 250
0.2
0.4
0.6
0.8
1
1.2
S/Lhole
f cr,h
oles
/f cr,n
o ho
les
hhole/h=0.66
hhole/h=0.44
hhole/h=0.19
hhole/h=0.26
2 3 4 50.75
0.8
0.85
0.9
S/Lhole
Figure 3.11 Influence of slotted hole spacing on the elastic buckling load of a long simply supported rectangular plate
31
When the holes are smaller relative to the plate width (hhole/h<0.44) and are spaced
closely together (S/ Lhole < 4), the local buckling influence of adjacent holes combine to
sharply decrease the elastic buckling stress. The inset of 930HFigure 3.11 highlights this
reduction in elastic buckling stress for hhole/h=0.19 and hhole/h=0.26, and 931HFigure 3.12
provides a summary of the associated buckled shapes. When hole spacing increases
beyond S/Lhole=5, the elastic buckling stresses approach constant magnitudes for all plate
widths considered, which is consistent with the trends presented in 932HFigure 3.6. This
observation is important from a design perspective because it serves as a rational basis
for setting hole spacing limits in cold‐formed steel members.
hhole/h=0.66, S/Lhole=4
hhole/h=0.44, S/Lhole=4
hhole/h=0.26, S/Lhole=4
Buckling is dampened at the holes, half-waves form between holes
Buckling of the unstiffened strips adjacent to the hole is dominant here
Buckled half-waves form along the length of the plate
Figure 3.12 Comparison of buckled shapes for a long stiffened element (L=24 Lhole ) with a slotted hole spacing of S/Lhole=4 and hhole/h=0.66, 0.44, and 0.26.
933HFigure 3.12 highlights the two types of buckling modes that can occur in stiffened
elements, plate buckling and unstiffened strip buckling. The influence of these buckling
32
modes on fcr is reflected in 934HFigure 3.13. The maximum decrease in fcr occurs for a
relatively small hole when compared to the plate width (hhole/h=0.30) and lies at the
transition between plate buckling, where axial stiffness of the buckled cells is reduced
with the presence of holes, and unstiffened strip buckling. Unstiffened strip buckling
occurs between hhole/h=0.30 and hhole/h=0.55 resulting in a relative increase in fcr as the strips
adjacent to the holes increase the axial stiffness of the plate. As hhole/h increases past
hhole/h=0.55 the unstiffened strip adjacent to the hole becomes narrow and stiff, resulting
in plate buckling away from the holes and an fcr similar to a plate without a hole. (An
increase in critical elastic buckling stress for large holes does not necessarily correspond
to an increase in ultimate strength because the strength of the plate will be limited by the
strength of the net cross‐section.) This is another important observation that will be
used when developing an elastic buckling prediction method for stiffened elements with
holes in Section 935H3.3.4.
33
0 0.2 0.4 0.6 0.8 10
0.5
1
1.5
hhole/h
f cr,h
oles
/f cr,n
o ho
les
Plate buckling Unstiffened strip buckling
Plate bucklingaway from the hole
Figure 3.13 Variation in fcr with increasing hhole/h for a stiffened element correspond to buckling mode shapes (see 936HFigure 3.12 for examples of plate buckling and unstiffened strip buckling mode shapes)
3.4 57BApproximate prediction method for use in design Approximations for the critical elastic buckling stress of stiffened elements (e.g.
column web or flange of a lipped C‐section) with holes under uniaxial compression are
developed in this section considering two elastic buckling states, buckling of the plate
without hole influence and buckling of the unstiffened strips adjacent to the hole. The
proposed prediction method is validated with thin shell finite element eigenbuckling
analyses for a variety of hole shapes, sizes, and spacings. Mandatory dimensional
tolerances on the prediction method are explicitly defined, and optional dimensional
limits, marked with an asterisk (*), are provided to avoid excessive conservatism.
34
3.4.1 118BDefinitions and assumptions
937HFigure 3.2 defines the plate and hole dimension notation used in the element
prediction method, including the hole spacing S, plate width h, and hole length and hole
width, Lhole and hhole. δhole is the offset distance of a hole measured from the centerline of
the plate. The elastic buckling prediction method for a stiffened element is developed
assuming a long plate loaded uniaxially and simply‐supported on all four sides with
evenly spaced holes. A summary of all prediction method equations is provided in
938HAppendix D.
3.4.2 119BPrediction Equations
The elastic buckling stress of a stiffened element with holes is approximated as
[ ]crhcrcr fff ,min=l . (3.1)
The critical elastic buckling stress for plate buckling (without hole influence) is
( )2
2
2
112⎟⎠⎞
⎜⎝⎛
−=
htEkfcr ν
π, (3.2)
where k is commonly taken equal to 4 when considering long rectangular plates (L/h>4).
When elastic buckling of the stiffened element is governed by the buckling of an
unstiffened strip adjacent to the hole, the critical elastic buckling stress of the governing
unstiffened strip is:
[ ]crBcrAnetcrh fff ,min, = (3.3)
35
( )2
2
2
112 ⎟⎟⎠
⎞⎜⎜⎝
⎛−
=i
icri htEkf
νπ and i = A or B (3.4)
The plate buckling coefficient ki for unstiffened strips A and B are approximated by (Yu
and Schafer 2007):
for 1≥ihole hL , ( ) 6.0
2.0425.0 95.0 −+=
iholei hL
k , (3.5)
for 1<ihole hL , 925.0=ik , and i = A or B. (3.6)
Eq. 939H(3.5) accounts for the length of the unstiffened strip. As hole length shortens relative
to the unstiffened strip width, ki increases. This is an improvement over AISI‐S100 which
conservatively assumes the lowerbound k=0.425 regardless of hole length. When Lhole/hi is
less than 1, k may be conservatively assumed equal to 0.925 via Eq. 940H(3.6) or calculated
directly by solving the classical stability equations for an unstiffened element
(Timoshenko 1961).
fcrh,net is the critical elastic buckling stress of the wider unstiffened strip
1P
2P
gcrhcr AfP =
⎟⎠⎞
⎜⎝⎛ −==
hhf
AAff hole
netcrhg
netnetcrhcrh 1,,
hhole h
fcrh,net
netnetcrhcr AfPPP ,21 ==+
( ) thhA holenet −= htAg =
Figure 3.14 Unstiffened strip elastic buckling stress conversion from the net to the gross section
36
To compare the buckling stress from the unstiffened strip (fcrh,net) to that of the entire plate
(fcr) equilibrium between the net and gross section must be considered, as shown in
941HFigure 3.14 and provided in the following:
( )hhff holenetcrhcrh −= 1, . (3.7)
3.4.3 120BVerification and equation limits
3.3.4.3.1 184BHoles centered transversely in plate
Thin shell finite element eigenbuckling analysis in ABAQUS, as described in Section
942H3.2, is employed here to verify the accuracy of the approximate prediction method in
Section 943H3.3.4.2. The boundary and loading conditions assumed for the stiffened element
are described in 944HFigure 3.5. The length of the slotted hole, Lhole, width of the plate h, the
shape of hole (slotted, circular, square), the hole spacing S, length of the plate L, and
plate thickness t are varied in the analyses. The plate and hole dimensions as well as the
ABAQUS critical elastic buckling stress, fcrl, for the 145 models considered, are provided
in 945HAppendix B(the eigenbuckling results from the studies in Section 946H3.3.2 and Section
947H3.3.3 are included in the 145 models). The parametric ranges in this study are
summarized for each hole type in 948HTable 3.2.
Table 3.2 Parameter ranges in stiffened element verification study. Hole type hhole/h S/Lhole S/h h/t # of models
Min 0.10 1.7 1.2 21Max 0.70 24.0 42.2 434Min 0.10 13.3 1.3 62Max 0.70 13.3 9.3 434Min 0.10 13.3 1.3 62Max 0.70 13.3 9.3 434
Slotted
Circular
Square
131
7
7
37
The results of the ABAQUS eigenbuckling analyses are compared to the stiffened
element prediction method in 949HFigure 3.15 and 950HFigure 3.16. 951HFigure 3.15 demonstrates that
as hole spacing S becomes small relative to the plate width h, the prediction method is
not always accurate. As hole spacing decreases, holes begin to coincide with the local
buckling half‐wavelengths (which have a length of h) and the influence of the individual
holes act cumulatively to decrease the axial stiffness of the plate. A similar loss in
stiffness is observed in 952HFigure 3.16 as hole spacing decreases relative to hole length.
From these observations, the following limits are imposed on the prediction method:
5.1≥hS
, (3.8)
2≥holeLS . (3.9)
If the parameter limit in Eq. 953H(3.9) is substituted into Eq. 954H(3.8), a third dimensional limit is
automatically imposed:
75.0≤h
Lhole (3.10)
Eq. 955H(3.10) prevents the hole length from being too long relative to the half‐wavelength of
the plate. The mean and standard deviation of the ABAQUS to predicted ratio for the
stiffened elements within the limits of Eq. 956H(3.8) and Eq. 957H(3.9) are 1.02 and 0.04
respectively, demonstrating that the prediction method is viable.
38
0 10 20 30 40 500
0.5
1
1.5
S/h
f crl ,
ABAQ
US/f cr
l , pr
edic
ted
Plate buckling controlsUnstif fened strip controls
0 10 20 30 40 50
0
0.5
1
1.5
S/h
f crl ,
ABAQ
US/f cr
l , pr
edic
ted
Plate buckling controlsUnstif fened strip controls
Figure 3.15 Accuracy of stiffened element prediction method as a function of hole spacing S to plate width h (a) without and (b) with the dimensional limits in Eq. 958H(3.8) and Eq.959H(3.9)
0 5 10 15 20 250
0.5
1
1.5
S/Lhole
f crl ,
ABAQ
US/f cr
l , pr
edic
ted
Plate buckling controlsUnstiffened strip controls
0 5 10 15 20 25
0
0.5
1
1.5
S/Lhole
f crl ,
ABAQ
US/f cr
l , pr
edic
ted
Plate buckling controlsUnstif fened strip controls
Figure 3.16 Accuracy of stiffened element prediction method as a function of hole spacing S to length of hole Lhole (a) without and (b) with the dimensional limits in Eq. 960H(3.8) and Eq.961H(3.9)
0 0.2 0.4 0.6 0.8 10
0.5
1
1.5
hhole/h
f crl ,
ABAQ
US/f cr
l , pr
edic
ted
Plate buckling controlsUnstiffened strip controls
0 0.2 0.4 0.6 0.8 1
0
0.5
1
1.5
hhole/h
f crl ,
ABAQ
US/f cr
l , pr
edic
ted
Plate buckling controlsUnstiffened strip controls
Figure 3.17 Accuracy of the stiffened element prediction method as a function of hole width hhole to plate
width h (a) without and (b) with the dimensional limits in Eq. 962H(3.8) and Eq.963H(3.9)
39
As hole width increases relative to plate width in 964HFigure 3.17, the controlling buckled
state transitions from buckling of the unstiffened strip adjacent to plate buckling. The
strips of web material adjacent to the holes have a higher axial stiffness than the sections
of the plate without holes, causing plate buckling to occur between the holes as shown
in 965HFigure 3.18.
Strips of plate adjacent to the hole are stiffer than plate between holes
Figure 3.18 For plates where the unstiffened strip is narrow compared to the plate width, plate buckling occurs between the holes.
As the hole width becomes small relative to plate width, the unstiffened strip
buckled state is predicted by the simplified method for slotted holes, although the actual
behavior is a combination of plate buckling and local buckling at the holes, as shown in
966HFigure 3.19. The assumption of unstiffened strip buckling when the slotted hole width is
small relative to plate width is conservative, with a maximum ABAQUS to predicted
ratio of 1.16 when hhole/h is in the range of 0.30. 967HFigure 3.19 also demonstrates that plate
buckling dominates over unstiffened strip buckling for stiffened elements with square
and circular holes. The prediction method identifies this elastic buckling behavior and
accurately predicts fcrl as shown in 968HFigure 3.20, where stiffened element results
containing just square or just circular holes are plotted.
40
Plate buckling and unstiffened strip buckling are both present when 0.20≤hhole/h≤0.60
Plate buckling dominates over unstiffened strip buckling for square (and circular holes)
Figure 3.19 Plate buckling and unstiffened strip buckling may both exist for a plate with holes. These modes are predicted conservatively as unstiffened strip buckling.
0 0.2 0.4 0.6 0.8 10
0.5
1
1.5
hhole/h
f crl,
ABAQ
US/f cr
l, pr
edic
ted
Plate buckling controls
Figure 3.20 Accuracy of prediction method for stiffened elements with square or circular holes as a function
of hole width hhole to plate width h.
3.3.4.3.2 185BOffset holes
43 additional ABAQUS eigenbuckling analyses were performed to evaluate the
accuracy of the simplified prediction method in Section 969H3.3.4.2, but now with
transversely offset holes. For these models, the hole offset from the centerline of the
41
plate, δhole, and the plate width, h, were varied. All plate models in this study have
regularly spaced slotted holes (S=20 in.) and constant plate length, L, of 100 in. The
boundary and loading conditions assumed for the stiffened element are described in
970HFigure 3.5. The model dimensions and critical elastic buckling stress, fcrl, for the 43
models considered, are summarized in 971HAppendix B. hstrip is the widest unstiffened strip,
either hA and hB. The parametric ranges for this study are summarized in 972HTable 3.3.
Table 3.3 Parameter range for stiffened element verification study with offset holes. Hole type hhole/h S/Lhole S/h h/t δhole/h # of models
Min 0.10 5.0 1.3 62 0.000Max 0.70 5.0 9.3 434 0.375
Slotted 43
0 0.2 0.4 0.6 0.8 10
0.5
1
1.5
2
2.5
3
hstrip/h
f crl ,
ABAQ
US/f cr
l , pr
edic
ted
Plate buckling controlsUnstif fened strip controls
0 0.2 0.4 0.6 0.8 1
0
0.5
1
1.5
2
2.5
3
hstrip/h
f crl ,
ABAQ
US/f cr
l , pr
edic
ted
Plate buckling controlsUnstif fened strip controls
Figure 3.21 Accuracy of the stiffened element elastic buckling prediction method as a function of
unstiffened strip width hstrip versus plate width h for offset holes (a) without and (b) with the dimensional
limits in Eq. 973H(3.8) and Eq.974H(3.9)
The ABAQUS critical elastic buckling stress results are compared to the prediction
method in 975HFigure 3.21, and demonstrate that the prediction method is conservative and
that the accuracy of the method improves as hstrip decreases relative to the plate width h
and hole length Lhole. The unstiffened strip buckled state is predicted to control for most
of the plate models, primarily because the shift in hole location results in a wider
42
unstiffened strip with less axial stiffness than that provided by the plate material
between holes. When the plate is relatively wide compared to the width of the hole and
the hole is shifted near the edge of the plate as shown in 976HFigure 3.22, the predictions can
be very conservative. The wide unstiffened strip is not a good approximation of the
actual behavior of the plate in this case. Prediction accuracy varies with hole offset, δhole,
as shown in 977HFigure 3.23a, and is most conservative as the hole offset becomes large
relative to the plate width h. To avoid overly conservative results, the following limit on
hole offset δhole is proposed for stiffened elements:
15.0≤hholeδ
* (3.11)
The mean and standard deviation of the ABAQUS to predicted ratio for the data within
the dimensional limits of Eq. 978H(3.8), Eq. 979H(3.9), and Eq. 980H(3.11) are 1.14 and 0.15 respectively
(also see 981HFigure 3.23b).
hstrip
The prediction method conservatively assumes unstiffened strip buckling of the wider strip adjacent to the hole, although plate buckling is observed.
Figure 3.22 Holes at the edge of a wide stiffened plate reduce the axial stiffness (and critical elastic buckling stress) but do not change the buckled shape.
43
0 0.1 0.2 0.3 0.4 0.50
0.5
1
1.5
2
2.5
3
δhole/h
f crl ,
ABAQ
US/f cr
l , pr
edic
ted
Plate buckling controlsUnstif fened strip controls
0 0.1 0.2 0.3 0.4 0.5
0
0.5
1
1.5
2
2.5
3
δhole/h
f crl,
ABAQ
US/f cr
l, pr
edic
ted
Plate buckling controlsUnstiffened strip controls
Figure 3.23 Accuracy of the stiffened element elastic buckling prediction method as a function of hole offset δhole versus plate width h for offset holes (a) without and (b) with the dimensional limits in Eq. 982H(3.8), Eq.983H(3.9),
and Eq. 984H(3.11)
3.4 29BStiffened element in bending
4.1 58BBoundary and loading conditions
The stiffened element is modeled with simply‐supported boundary conditions and
loaded with a bending compressive stress distribution as shown in 985HFigure 3.24. The
location of the neutral axis about which bending occurs, Y, is measured from the
compressed edge of the plate.
1
2
3
Restrain plate perimeter in 2 (v=0)
Restrain transverse midline in 1 (u=0)
Restrain point in 3 (w=0)
Restrain point in 3 (w=0)
Y
Neutral axis
Figure 3.24 Boundary and loading conditions for a stiffened element in bending
44
4.2 59BInfluence of transversely-centered slotted holes Shell finite element eigenbuckling models of stiffened elements with regularly
spaced slotted holes are evaluated in this study. The bending stress distribution is
symmetric about the transverse centerline of the plate (Y=0.50h) for all models. The
slotted holes are centered transversely in the plate (δhole=0). The plate and hole
dimensions and the critical elastic buckling stress, fcrl, for the 28 models considered, are
summarized in 986HAppendix B. The parametric ranges for this study are summarized in
987HTable 3.4.
Table 3.4 Parameter ranges considered for stiffened elements in bending with holes. Hole type hhole/h S/Lhole S/h h/t Y/h # of models
Min 0.10 1.67 1.33 61.93 0.50Max 0.70 5.00 9.33 433.53 0.50
28Slotted
988HFigure 3.25 highlights the influence of hole width to plate width on stiffened
elements in bending. As hhole/h increases, the buckling mode transitions from plate
buckling (similar to a plate without a hole) to buckling of the compressed unstiffened
strip adjacent to the hole. The buckled half‐wavelength of a plate in bending is between
0.25h to 0.50h, which results in a shortened half‐wavelength of the unstiffened strip
(often less than the length of the hole) when compared to the equivalent unstiffened
strip buckling mode for stiffened elements in uniaxial compression (See Section 989H3.3).
45
Unstiffened strip buckling becomes more predominant as the hole size increases relative to plate width.
Unstiffened strip half-wavelength can be less than the length of the hole.
hhole/h=0.10 hhole/h=0.30 hhole/h=0.50
Figure 3.25 Stiffened plates loaded with a linear bending stress gradient exhibit buckling of the unstiffened strip adjacent to the hole in the compression region of the plate.
The maximum reduction in critical elastic buckling stress occurs in the range of
hhole/h=0.30 as shown in 990HFigure 3.26a. This result is consistent with the elastic buckling
results for stiffened plates under axial compression (See 991HFigure 3.13). The elastic
buckling behavior of stiffened elements in bending are different than in pure
compression though as hhole/h exceeds 0.50. Unstiffened strip buckling continues to
dominate for plate bending (with an associated reduction in fcr) while plate buckling
away from the hole controls for uniaxially compressed plates (with minimal influence
on fcr even for very large holes). This distinction between compression (columns) and
bending (beams) elastic buckling behavior of stiffened elements is important when
considering how to approximate elastic buckling behavior. fcr decreases as hole spacing
becomes small relative to hole length as shown in 992HFigure 3.26b, identifying S/Lhole as
another important parameter when predicting elastic buckling of stiffened elements in
bending (as it is for stiffened elements in compression, See Section 993H3.3).
46
0 0.2 0.4 0.6 0.8 10
0.5
1
1.5
hhole/h
f cr,h
ole/f cr
,no
hole
0 2 4 6 8 100
0.5
1
1.5
S/Lhole
f cr,h
ole/f cr
,no
hole
Figure 3.26 Influence of slotted holes on critical elastic buckling stress fcr of stiffened elements in bending as a function of (a) hole size relative to plate width and (b) hole spacing as a function of hole length.
4.3 60BInfluence of offset slotted holes
4.3.1 121BNeutral axis location at Y=0.50h
Shell finite element eigenbuckling models of stiffened elements with regularly
spaced offset slotted holes are evaluated in this study. The bending stress distribution is
symmetric about the transverse centerline of the plate (Y=0.50h) for all models. The hole
offset, δhole, ranges from ‐0.375h to +0.375h, where a positive shift moves the holes into the
compression region of the plate. The plate and hole dimensions and the critical elastic
buckling stress, fcr, for the 92 models considered, are summarized in 994HAppendix B. The
parameter range considered in this study is provided in 995HTable 3.5.
Table 3.5 Study parameter limits for stiffened element in bending (Y/h=0.50) with offset holes
Hole type hhole/h S/Lhole S/h h/t Y/h δhole/h # of modelsMin 0.10 5.00 1.33 61.93 0.50 -0.375Max 0.70 5.00 9.33 433.53 0.50 0.375
92Slotted
The presence of holes in the compression region of a stiffened element in bending
(Y=h/2) decreases the critical elastic buckling stress when compared to a plate without
holes as shown in 996HFigure 3.27. Depending upon the width of the unstiffened strip “A”
47
in the compressed region of the plate and the unstiffened strip “B” in the tensile region
of the plate (see 997HFigure 3.3 for definitions) relative to hole depth h, unstiffened strip
buckling may occur above the hole, below the hole, or above and below the hole. fcr
varies with the transverse position of the holes in the plate (characterized as the width of
unstiffened strip “A”, hA) in 998HFigure 3.27. The trends in fcr can be related to the elastic
buckling modes in 999HFigure 3.28. If the holes are located in the tensile region of the
stiffened element, the buckled mode shape (and fcr) are unchanged when compared to a
stiffened element without holes. The relationship between these buckled mode shapes
and trends in fcr will be used in Section 1000H3.4.4 when developing an approximate elastic
buckling prediction method for stiffened elements in bending.
0 0.5 1 1.5
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1
fcr,hole/fcr,no hole
h A/h
hhole/h=0.10
hhole/h=0.20
hhole/h=0.30
hhole/h=0.40
hhole/h=0.50
hhole/h=0.60
hhole/h=0.70
Neutral axis
Compression
Tension
A
B
C
D
Figure 3.27 Hole location influence on critical elastic buckling stress fcr for a stiffened plate in bending (Y=0.50h) (Buckled mode shapes corresponding to A, B, C, and D are provided in 1001HFigure 3.28.)
48
AB
C D
Unstiffened strip buckling (below hole) Unstiffened strip
buckling (below and above hole)
Unstiffened strip buckling (above hole)
Plate buckling (no hole influence)
hA
Figure 3.28 The buckled mode shape changes as slotted holes move from the compression region to the tension region of a stiffened element in bending (hhole/h=0.20).
4.3.2 122BNeutral axis location at Y=0.75h
The neutral axis in the shell finite element eigenbuckling models from Section 1002H3.4.3.2
is now modified to Y=0.75h. The trends in fcr in 1003HFigure 3.29 are similar to those observed
in 1004HFigure 3.27 (Y=0.50h). Elastic buckling of the unstiffened strip below the holes occurs
when the hole is close to the compressed edge. The mode shape transitions to
unstiffened strip buckling above the holes as the hole offset increases toward the tensile
region of the plate. The plate and hole dimensions and the critical elastic buckling
stress, fcr, for the 92 models considered here, are summarized in 1005HAppendix B.
49
0 0.5 1 1.5
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1
fcr,hole/fcr,no hole
h A/h
hhole/h=0.10
hhole/h=0.20
hhole/h=0.30
hhole/h=0.40
hhole/h=0.50
hhole/h=0.60
hhole/h=0.70
Neutral axis
Compression
Tension
Figure 3.29 Hole location influence on critical elastic buckling stress fcr for a stiffened plate in bending (Y=0.75h)
4.4 61BApproximate prediction method for use in design In the previous section unique elastic buckling modes were identified for a stiffened
element in bending with holes. Buckling of the unstiffened strip between the hole and
the compressed edge of the plate (unstiffened strip “A”) or between the hole and the
tension edge of the plate (unstiffened strip “B”) may occur depending upon the
transverse location of the hole in the plate, the width of the hole (hhole) relative to the
depth of the plate (h), and the location of the plate neutral axis (Y). If the hole is
completely contained within the tension region of the plate then the hole has a minimal
influence on elastic buckling and the critical elastic buckling stress, fcr, remains
unchanged. These observations can be used to define an approximation for the critical
elastic buckling stress of a stiffened element with holes in bending:
50
[ ]crhcrcr fff ,min=l . (3.12)
The critical elastic buckling stress for a stiffened element in bending (without the
influence of holes), fcr, may be determined with Eq. 1006H(3.2) where the buckling coefficient k
is calculated with AISI‐S100‐07 Eq. B2.3‐2 (AISI‐S100 2007):
( ) ( )ψψ ++++= 12124 3k (3.13)
and ψ is the absolute value of the ratio of tensile stress to compressive stress applied to
the stiffened element, i.e.:
( ) YYhff −== 12ψ . (3.14)
When elastic buckling of the stiffened element is governed by the buckling of an
unstiffened strip adjacent to a hole, the critical elastic buckling stress is:
[ ]crBcrAnetcrh fff ,min, = (3.15)
Consideration of unstiffened strip “A” is required only if hA<Y, i.e., at least a portion of
the hole must lie in the compression region of the stiffened element. If that condition is
met the elastic buckling stress for strip “A” is:
( )2
2
2
112 ⎟⎟⎠
⎞⎜⎜⎝
⎛−
=A
AcrA htEkf
νπ (3.16)
The plate buckling coefficient for the unstiffened strip “A” is approximated as
( )2035.0024.0
76.170.234.0
578.0
AholeA
A
AA hL
k++
−+
+=
ψψ
ψ, and
YhY A
A−
=ψ (3.17)
Eq. 1007H(3.17) is a modification of AISI‐S100‐07 Eq. B3.3‐2 (AISI‐S100 2007) This expression
accounts for the gradient of the compressive stress distribution and the aspect ratio of
51
the unstiffened strip (see 1008HAppendix C for derivation). The equation for ψA is derived in
1009HFigure 3.30.
f2
f1
hA
Y
Similar Triangles
Neutral Axis
YhY
ff A
A−
==1
2ψ
Figure 3.30 Derivation of stress ratio ψΑ for unstiffened strip “A”.
Consideration of unstiffened strip “B” is required only if hA+hhole<Y, i.e., only when
the entire hole lies within the compressed region of the plate. For this case the buckling
stress of the unstiffened strip, converted to a stress at the compressed edge, is found as:
⎟⎟⎠
⎞⎜⎜⎝
⎛−−⎟⎟
⎠
⎞⎜⎜⎝
⎛−
=holeAB
BcrB hhYY
htEkf
2
2
2
)1(12 νπ , (3.18)
where the final term in Eq. 1010H(3.18) converts the buckling stress from the edge of
unstiffened strip “B” to the edge of unstiffened strip “A” as shown in 1011HFigure 3.31 so that
the two stresses (fcrA and fcrB) may be compared in Eq. 1012H(3.15) to determine the minimum.
The plate buckling coefficient for the unstiffened strip “B” is approximated as:
for Lhole/hB>2
573.0100.0340.0 2 ++= BBBk ψψ , (3.19)
for Lhole/hB≤2
52
14.020.0
49.06.138.0
1.03.0
28.1
+⎟⎟⎠
⎞⎜⎜⎝
⎛+−
+⎟⎟⎠
⎞⎜⎜⎝
⎛+
=
hole
BB
hole
BB
B
Lh
Lh
k
ψ
ψ, (3.20)
and the ratio of tension to compressive stresses (derived in 1013HFigure 3.31) is:
holeA
B hhYYh−−
−=ψ , 100 ≤≤ Bψ . (3.21)
The plate buckling coefficient kB is applicable over a larger range of ψB than AISI‐S100‐07
Eq. B3.2‐5 (AISI‐S100 2007) and accounts for the increase in kB as the unstiffened strip
aspect ratio tends to zero (i.e., a wide, short strip resulting from a small hole). Refer to
1014HAppendix C for the derivation of kB.
f2
f1
hA
Y
Similar Triangles
Neutral AxisfcrB
h
holeAB hhY
Yhff
−−−
==1
2ψ
holeA
crB
hhYY
ff
−−=
1
1fhhYYf
holeAcrhB −−
=
Similar Triangles
Solve for fcrB
Figure 3.31 Derivation ψB and conversion of the compressive stress at the edge of unstiffened strip “B” to the stress fcrB at the edge of the plate
Conversion to the gross section for the comparison of stresses in Eq. 1015H(3.12) requires
that:
for hA+hhole ≥ Y, ( )Yhff A
Anetcrhcrh ψ+= 1, , (3.22)
53
for hA+hhole < Y, ⎥⎦
⎤⎢⎣
⎡⎟⎠⎞
⎜⎝⎛ −−=
Yh
Yhff hole
Ahole
netcrhcrh ψ21, . (3.23)
The conversion from fcrh,net at the net section of the plate to fcrh on the gross cross‐section is
obtained with a similar method to that described in 1016HFigure 3.14 for stiffened elements in
uniaxial compression; the total compressive force at the net and gross cross‐sections are
assumed in equilibrium as shown in 1017HFigure 3.32 and 1018HFigure 3.33. A summary of all
prediction method equations is provided in 1019HAppendix D.
Neutral Axis
fcrh,net fcrhY
hA
Free Body Diagram
PP
Force Equilibrium
Solve for fcrh
ψΑfcrh,net
Ytfthff
P crhAnetcrhAnetcrh
21
2,, =⎟⎟
⎠
⎞⎜⎜⎝
⎛ +=
ψ
( )Yhff A
Anetcrhcrh ψ+= 1,
Figure 3.32 Derivation of fcrh for the case when hA+hhole≥Y (when the hole is located partially in the compressed region and partially in the tension region of the plate)
54
Neutral Axis
fcrh,net fcrhY
hA
Free Body Diagram
P PψΑfcrh,net
hhole
f3
Ytfthff
tYfP crhholenetcrhA
netcr 21
221 3,
, =⎟⎟⎠
⎞⎜⎜⎝
⎛ +−=
ψ
Yh
YhhY
ff hole
AholeA
netcrh
−=−−
= ψ,
3
⎟⎟⎠
⎞⎜⎜⎝
⎛⎟⎠⎞
⎜⎝⎛ −+−= netcrh
holeAnetcrhA
holenetcrhcrh f
Yhf
Yhff ,,, ψψ
⎥⎦
⎤⎢⎣
⎡⎟⎠⎞
⎜⎝⎛ −−=
Yh
Yhff hole
Ahole
netcrhcrh ψ21,
Force Equilibrium
Define f3 using similar triangles
Substitute f3 and solve for fcrh
Simplify
Figure 3.33 Derivation of fcrh for the case when hA+hhole<Y (hole lies completely in the compressed region of the plate).
4.5 62BVerification and parameter limits
The elastic buckling prediction method for stiffened elements in bending is now
evaluated with the ABAQUS eigenbuckling results presented in Section 1020H3.4.2 and
Section 1021H3.4.3. The viability of the method is examined for evenly spaced slotted holes
centered transversely or offset in a plate. Parameter limits on the prediction method,
required when formalizing the method for use in design, are also identified.
ABAQUS results are compared to predictions in 1022HFigure 3.34a and 1023HFigure 3.38a.
1024HFigure 3.34a demonstrates that the simplified method underpredicts the elastic buckling
stress as aspect ratio of the unstiffened strip “A” increases. A dimensional tolerance is
imposed to avoid unconservative predictions:
55
10≤A
hole
hL
. (3.24)
Eq. 1025H(3.24) also serves as a practical limit on the slenderness of an unstiffened strip, and
therefore is also is imposed on the unstiffened strip “B”:
10≤B
hole
hL
. (3.25)
The prediction method becomes increasingly conservative as hA/Y approaches unity as
shown in 1026HFigure 3.35a. When only a small portion of the hole exists in the compressed
region of the plate, the observed buckling mode is more consistent with plate buckling
than unstiffened strip buckling as predicted by the simplified method (See 1027HFigure 3.28,
picture D). A dimensional limit is suggested to prevent excessive conservatism in this
case:
6.0≤YhA * (3.26)
The hole spacing limits defined in Section 1028H3.3.4.3 for stiffened elements in uniaxial
compression are also considered here for a stiffened element in bending. The prediction
accuracy degrades when hole spacing S approaches the plate width h as shown in 1029HFigure
3.36a. Predictions can also be unconservative when S is 2 to 3 times the length Lhole as
shown in 1030HFigure 3.37a. With the limits from Eq. 1031H(3.8), Eq. 1032H(3.9), Eq. 1033H(3.24), Eq. 1034H(3.25), and
Eq. 1035H(3.26) imposed, the method is observed to be viable predictor over a wide range of
hhole/h as shown in 1036HFigure 3.38b. The mean and standard deviation of the ABAQUS to
predicted ratio within the imposed dimensional limits are 1.22 and 0.11 respectively.
56
0 5 10 15 20 25 30 35 400
0.5
1
1.5
2
2.5
3
Lhole/hA
f cr,A
BAQ
US/f cr
,pre
dict
ed
plate bucklingunstiffened strip "A"unstiffened strip "B"
0 5 10 15 20 25 30 35 400
0.5
1
1.5
2
2.5
3
Lhole/hA
f cr,A
BAQ
US/f cr
,pre
dict
ed
plate bucklingunstiffened strip "A"unstiffened strip "B"
Figure 3.34 Influence of Lhole/yA on the accuracy of the prediction method for stiffened elements in bending
(a) without and (b) with the dimensional limits defined in Eq. 1037H(3.9), Eq. 1038H(3.24), Eq. 1039H(3.25), and Eq. 1040H(3.26).
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 10
0.5
1
1.5
2
2.5
3
hA/Y
f cr,A
BAQ
US/f cr
,pre
dict
ed
plate bucklingunstiffened strip "A"unstiffened strip "B"
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 10
0.5
1
1.5
2
2.5
3
hA/Y
f cr,A
BAQ
US/f cr
,pre
dict
ed
plate bucklingunstiffened strip "A"unstiffened strip "B"
Figure 3.35 Influence of hA/Y on the accuracy of the prediction method for stiffened elements in bending (a)
without and (b) with the dimensional limits defined in Eq. 1041H(3.9), Eq. 1042H(3.24), Eq. 1043H(3.25), and Eq. 1044H(3.26).
57
0 1 2 3 4 5 6 7 8 9 100
0.5
1
1.5
2
2.5
3
S/h
f cr,A
BAQ
US/f cr
,pre
dict
ed
plate bucklingunstiffened strip "A"unstiffened strip "B"
0 1 2 3 4 5 6 7 8 9 10
0
0.5
1
1.5
2
2.5
3
S/h
f cr,A
BAQ
US/f cr
,pre
dict
ed
plate bucklingunstiffened strip "A"unstiffened strip "B"
Figure 3.36 Influence of S/h on the accuracy of the prediction method for stiffened elements in bending (a) without and (b) with the dimensional limits defined in Eq. 1045H(3.9), Eq. 1046H(3.24), Eq. 1047H(3.25), and Eq. 1048H(3.26).
0 1 2 3 4 5 6 7 8 9 100
0.5
1
1.5
2
2.5
3
S/Lhole
f cr,A
BAQ
US/f cr
,pre
dict
ed
plate bucklingunstiffened strip "A"unstiffened strip "B"
0 1 2 3 4 5 6 7 8 9 100
0.5
1
1.5
2
2.5
3
S/Lhole
f cr,A
BAQ
US/f cr
,pre
dict
ed
plate bucklingunstiffened strip "A"unstiffened strip "B"
Figure 3.37 Influence of S/Lhole on the accuracy of the prediction method for stiffened elements in bending (a)
without and (b) with the dimensional limits defined in Eq. 1049H(3.9), Eq. 1050H(3.24), Eq. 1051H(3.25), and Eq. 1052H(3.26).
58
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 10
0.5
1
1.5
2
2.5
3
hhole/h
f cr,A
BAQ
US/f cr
,pre
dict
ed
plate bucklingunstiffened strip "A"unstiffened strip "B"
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 10
0.5
1
1.5
2
2.5
3
hhole/h
f cr,A
BAQ
US/f cr
,pre
dict
ed
plate bucklingunstiffened strip "A"unstiffened strip "B"
Figure 3.38 Influence of h/hhole on the accuracy of the prediction method for stiffened elements in bending (a)
without and (b) with the dimensional limits defined in Eq. 1053H(3.9), Eq. 1054H(3.24), Eq. 1055H(3.25), and Eq. 1056H(3.26).
3.5 30BUnstiffened element in uniaxial compression
5.1 63BBoundary and loading conditions
The unstiffened element is modeled with simply‐supported boundary conditions
on three sides and unsupported on the fourth side parallel to the application of a
uniform compressive stress as shown in 1057HFigure 3.39.
1
2
3
Restrain plate perimeter in 2 (v=0)
Restrain longitudinal midline in 3 (w=0)
Restrain transverse midline in 1 (u=0)
Figure 3.39 ABAQUS boundary and loading conditions for unstiffened plate loaded uniaxially.
59
5.2 64BInfluence of regularly-spaced holes Eigenbuckling analyses in ABAQUS are performed to evaluate the influence of
evenly‐spaced holes on the elastic buckling behavior of an unstiffened element. The
model loading and boundary conditions are summarized in 1058HFigure 3.39 and the material
properties and meshing procedures are the same as those described in Section 1059H3.2. The
plate width h, hole length Lhole, and hole type (slotted, circular, rectangular) are varied in
this study. The hole width remains constant at hhole=1.5 in. The plate and hole
dimensions as well as the critical elastic buckling stress, fcrl, for the 91 models considered,
are provided in 1060HAppendix B. The parametric ranges considered in this study for each
hole type are summarized in 1061HTable 3.6.
Table 3.6 Parameter range for study of regularly‐spaced holes on unstiffened elements. Hole Type hhole/h S/Lhole S/h h/t # of models
Min 0.10 1.7 1.0 21Max 0.70 24.0 42.2 434Min 0.10 13.3 1.3 62Max 0.70 13.3 9.3 434Min 0.10 13.3 1.3 62Max 0.70 13.3 9.3 434
Square
77
7
7
Slotted
Circular
A comparison of the ABAQUS results from the 91 models to the theoretical elastic
buckling stress for a long unstiffened element (k=0.425) in 1062HFigure 3.40 demonstrates that
the critical elastic buckling stress fcr decreases as hole width hhole increases relative to plate
width h. Holes always reduce the critical elastic buckling stress of unstiffened elements
in the cases studied. Buckling of the unstiffened strip “A” between the hole and the
simply supported edge is not observed in the simulations because L/h is always greater
than Lhole/hA, although buckling of the unsupported strip “B” at the free edge occurs as the
strip becomes slender (similar to Euler buckling) as shown in 1063HFigure 3.41. These
60
important observations are employed in Section 1064H3.5.4 to develop an approximate
prediction method for the critical elastic buckling stress of an unstiffened element with
holes.
0 0.2 0.4 0.6 0.8 10
0.2
0.4
0.6
0.8
1
1.2
1.4
hhole/h
f cr, h
ole/f cr
, no
hole
Figure 3.40 The presence of holes causes a decrease in critical elastic buckling load for unstiffened plates in uniaxial compression.
Holes do not influence the buckled shape of unstiffened plates until the hole width becomes large relative to plate width.
Buckling of the strip at the free edge of the plate changes the shape of the local buckling mode.
hhole/h=0.10 hhole/h=0.60
hA
Figure 3.41 Buckled shapes of unstiffened plates with holes.
61
5.3 65BInfluence of offset slotted holes ABAQUS eigenbuckling analyses were performed to evaluate the influence of
transversely offset slotted holes on the elastic buckling of an unstiffened element. The
ratio of transverse offset, δhole, to plate width h was varied from ‐0.375 to 0.375, where a
negative offset shifts the holes toward the simply supported edge and a positive offset
shifts towards the free plate edge (refer to 1065HFigure 3.2 for a definition of δhole). The model
loading and boundary conditions are summarized in 1066HFigure 3.39 and the material
properties and meshing procedures are the same as those described in Section 1067H3.2. The
plate and hole dimensions as well as the critical elastic buckling stress, fcrl, for the 92
models considered, are summarized in 1068HAppendix B. The parametric ranges considered
here are provided in 1069HTable 3.7.
Table 3.7 Parameter range considered for unstiffened element study with offset holes hhole/h S/Lhole S/h h/t δhole/h # of models
Min 0.10 5.00 1.33 62 -0.375Max 0.70 5.00 9.33 434 0.375
Slotted 92
The axial stiffness of an unstiffened element is higher near the simply supported
edge and lower near the free edge. It is hypothesized that holes shifted towards the
simply‐supported edge will reduce the critical elastic buckling stress more than hole
material removed from near the free edge. This hypothesis is confirmed in 1070HFigure 3.43
where fcr decreases more when holes are shifted towards the simply‐supported edge.
The dimension of the plate strip between the hole and the simply‐supported edge, hA
(see 1071HFigure 3.3), is identified as a useful parameter when predicting fcr. fcr forms a trend
line when plotted against Lhole relative to yA as demonstrated in 1072HFigure 3.43a for offset
62
holes. The same plot is produced using the data from Section 1073H3.5.2 for centered holes in
1074HFigure 3.43b with similar results. This important conclusion, that yA and Lhole are key
parameters influencing fcr, is used in the next section to develop an approximate elastic
buckling prediction method for unstiffened elements with holes.
-0.4 -0.3 -0.2 -0.1 0 0.1 0.2 0.3 0.40
0.2
0.4
0.6
0.8
1
1.2
1.4
δhole/h
f cr, h
ole/f cr
, no
hole
Data groups correspond to hhole/h=0.10, 0.20, …0.70
Figure 3.42 The critical elastic buckling stress of a stiffened plate decreases as holes are shifted toward the
simply supported edge (+δhole)
0 5 10 15 20 25 30 35 400
0.2
0.4
0.6
0.8
1
1.2
1.4
Lhole/hA
f cr, h
ole/f cr
, no
hole
0 5 10 15 20 25 30 35 400
0.2
0.4
0.6
0.8
1
1.2
1.4
Lhole/hA
f cr, h
ole/f cr
, no
hole
Figure 3.43 The critical elastic buckling stress for stiffened elements with (a) transversely offset holes and
(b) centered holes (from Section 1075H3.5.2) decreases as a function of hole length Lhole to hA
63
5.4 66BApproximate prediction method for use in design An approximate elastic buckling prediction method for an unstiffened element with
holes is presented here. The method is based on the observations in Section 1076H3.3.2 and
Section 1077H3.3.3 for long unstiffened elements with evenly spaced holes. The width of the
strip between the hole and the simply supported edge, hA, and the length of the hole Lhole
are utilized as predictors of the critical elastic buckling stress. A summary of the
prediction method equations are provided in 1078HAppendix D.
5.4.1 123BDerivation of empirical buckling coefficient
An empirical plate buckling coefficient is determined using a linear regression
analysis of the data in 1079HFigure 3.43a and 1080HFigure 3.43b for both centered and offset slotted
holes, which was then adjusted to have a slightly conservative bias. The regression
minimizes the error between the ABAQUS results and the classical stability solution of
an unstiffened element (k=0.425) for the plate models within the following parametric
limits:
10≤A
hole
hL (3.27)
10≤B
hole
hL (3.28)
50.0≤h
hhole (3.29)
Eq. 1081H(3.27) is imposed as a practical limit on the slenderness of the strip adjacent to the
hole at the simply‐supported plate edge. Eq. 1082H(3.28) prevents Euler buckling of
64
unstiffened strip “B” as shown in 1083HFigure 3.41. Eq. 1084H(3.29) is imposed because of the
increased rate of degradation in fcr observed in 1085HFigure 3.40 as holes become large relative
to plate width. The empirical plate buckling coefficient is set as:
⎟⎟⎠
⎞⎜⎜⎝
⎛−=
A
hole
hLk 062.01425.0 (3.30)
where the strip of plate between the hole and the simply supported edge, hA, is
calculated as
holehole
Ahhh δ−
−=
2. (3.31)
A positive δhole (hole offset from the centerline of the plate, See 1086HFigure 3.2) shifts the hole
towards the simply supported edge. The empirical buckling coefficient in Eq. 1087H(3.30) is
shown in 1088HFigure 3.44a to be a slightly conservative but accurate representation of
ABAQUS predicted buckling coefficients. The mean and standard deviation of the
ABAQUS to empirical prediction ratio are 1.06 and 0.09 respectively.
0 1 2 3 4 5 6 7 8 9 100
0.05
0.1
0.15
0.2
0.25
0.3
0.35
0.4
0.45
0.5
Lhole/yA
buck
ling
coef
f. k
ABAQUSEq. (3.30)
0 1 2 3 4 5 6 7 8 9 10
0
0.2
0.4
0.6
0.8
1
1.2
1.4
Lhole/yA
f cr, A
BAQ
US/f cr
, pre
dict
ed
Figure 3.44 (a) Comparison of ABAQUS and empirical plate buckling coefficients for an unstiffened element with holes and (b) ABAQUS to predicted elastic buckling stress for an unstiffened element
65
5.4.2 124BPrediction equations
The elastic buckling stress of an unstiffened element in compression with holes is
thus approximated as:
[ ]crhcrcr fff ,min=l . (3.32)
The critical elastic buckling stress prediction for plate buckling of the unstiffened
element without holes (fcr) is calculated with Eq. 1089H(3.2), where k=0.425 when considering
long rectangular plates (L/h>4). The minimum critical elastic buckling stress of the
unstiffened element with holes, fcrh, coincides with either buckling of the entire
unstiffened element with holes or buckling of the unstiffened strip “A” adjacent to the
hole and the simply supported edge, or:
( ) ⎥⎥⎦
⎤
⎢⎢⎣
⎡⎟⎠⎞
⎜⎝⎛ −⎟
⎠⎞
⎜⎝⎛
−=
hhf
htEkf hole
crAcrh 1,112
min2
2
2
νπ
(3.33)
where k is an empirical plate buckling coefficient derived from finite element
eigenbuckling studies in Eq. 1090H(3.30). fcrA is calculated with Eq. 1091H(3.4) and modified by the
factor (1‐ hhole/h) to convert the stress on the unstiffened strip “A” to the stress at the end
of the plate so that it can be compared to the buckling stress of the unstiffened element.
fcrh will always be predicted as less than or equal to fcr with this method.
66
Chapter 4 3BElastic buckling of cold-formed steel members with holes
The elastic buckling properties of cold‐formed steel lipped C‐section beams and
columns with holes are evaluated in this chapter using thin‐shell finite element
eigenbuckling analyses in ABAQUS. The elastic buckling studies are used to assess the
influence of holes on the local, distortional, and global critical elastic buckling loads Pcrl,
Pcrd, Pcre. The studies also identify elastic buckling modes unique to cold‐formed steel
members with holes. Elastic buckling properties of existing experiments on cold‐formed
steel columns and beams with holes are summarized and formal buckling modes are
defined in preparation for the presentation of the Direct Strength Method for structural
members with holes in 1092HChapter 8.
67
4.1 31BFinite element modeling assumptions
The elastic buckling behavior of the cold‐formed steel structural members with holes
are obtained with eigenbuckling analyses in ABAQUS (ABAQUS 2007a). All members
are modeled with ABAQUS S9R5 reduced integration nine‐node thin shell elements.
The typical finite element aspect ratio is 1:1 and the maximum aspect ratio is limited to
8:1 (refer to 1093HChapter 2 for a discussion on ABAQUS thin shell finite element types and
finite element aspect ratio limits). Element meshing is performed with a Matlab
(Mathworks 2007) program written by the author (refer to 1094HAppendix A for a description
of the program). Cold‐formed steel material properties are assumed as E=29500 ksi and
ν=0.3 in the finite element models unless noted otherwise. Py, the squash load of the
column, is calculated by multiplying an assumed yield stress of 50 ksi by the gross cross‐
sectional area of the column.
4.2 32BElastic buckling of columns with holes
2.1 67BMember and hole dimensions
Member and hole dimension notation used throughout this chapter is summarized
in 1095HFigure 4.1. Uppercase dimensions (H, D, B) are out‐to‐out and lowercase dimensions
(b, h) are flat lengths between points of curvature.
68
B1
H
D1
b
h
r
bhole
hhole
t
D2
B2R
Figure 4.1 C‐section and hole dimension notation
2.2 68BLoading and boundary conditions
The cold‐formed steel column boundary conditions are modeled as warping free at
the member ends and warping fixed at the midlength of the member as shown in 1096HFigure
4.2, which mimics the semi‐analytical finite strip method (Schafer and Ádàny 2006). The
columns are loaded at each end with stress distributions applied as consistent nodal
loads in ABAQUS (see Section 1097H3.2 for details on the loading implementation).
69
end nodes are supported in 2 and 3 (v = w = 0)
end nodes are supported in 2 and 3 (v = w = 0)
midspan nodes are supported in 1 (u = 0)
1
2
3
Figure 4.2 Columns are modeled with pinned warping‐free boundary conditions and compressed from both ends
2.3 69BElastic buckling comparison of short C-section columns versus
isolated stiffened elements
This study builds on the results and observations in 1098HChapter 3 for cross‐sectional
elements with holes and marks a transition in research focus from elements to full cold‐
formed steel members. The influence of one slotted hole on the elastic buckling
behavior of a range of rectangular plates and SSMA cold‐formed steel structural stud
sections is compared, the goal being to quantify the relative influence of a web hole on
one element in a cross‐section (in this case a stiffened element, see 1099HFigure 3.1 for
definition and 1100HFigure 3.5 for ABAQUS boundary conditions) versus a full C‐section. The
slotted hole has dimensions of hhole=1.5 in., Lhole=4 in., and rhole=0.75 in. The plate widths
are chosen to correspond with the flat web widths of standard SSMA structural studs
(SSMA 2001). Plate aspect ratios are held constant at 4:1. From each plate, a full
structural stud finite element model is developed for comparison. The SSMA member
70
designations and cross section dimensions considered in this study are listed in 1101HTable
4.1.
Table 4.1 SSMA structural stud and plate dimensions SSMA H B D r t h b hhole/h L=4h
Designation in. in. in. in. in. in. in. in.250S162-33 2.50 1.63 0.50 0.0764 0.0346 2.28 1.40 0.66 9.1350S162-33 3.50 3.28 0.46 13.1362S162-33 3.62 3.40 0.44 13.6400S162-33 4.00 3.78 0.40 15.1550S162-33 5.50 5.28 0.28 21.1600S162-33 6.00 5.78 0.26 23.1800S162-33 8.00 7.78 0.19 31.1
Before examining the elastic buckling load, consider the observed changes in the first
mode shape caused by the addition of a hole as given in 1102HFigure 4.3. For the buckled
shape of the SSMA 250S162‐33 in 1103HFigure 4.3a, the number of buckled half‐waves changes
from four to three for the isolated plate and from five to two for the full member, when
the hole is added. The strips of plate adjacent to the hole are stiffened by the connected
flange in the full member, causing buckled half‐waves to form in the web away from the
hole. Also, the length of the hole, Lhole, is approximately half of the length of the member
L in the SSMA 250S162‐33 member which also prevents local buckling in the web. In
1104HFigure 4.3b, the hole decreases the number of buckled half‐waves from four to three in
the SSMA 440S162‐33 isolated plate but does not change the number of half‐waves in the
full member. The cross‐section connectivity of the full member limits deformation at the
hole and encourages buckling half‐waves to form along the entire member. Also, there
is more web material to accommodate local buckling along the length (Lhole/L=0.26) when
compared to the SSMA 250S162‐33 member.
71
SSMA 250S162-33 hhole/h=0.66
SSMA 400S162-33 hhole/h=0.40
# of Local Buckling Half-Waves (Typ.)
3
4
2
5
3
4
5
5
(a) (b)
Figure 4.3(a) SSMA 250S162‐33 web plate and structural stud, and (b) SSMA 400S162‐33 web plate and
structural stud
1105HFigure 4.4 presents the influence of a slotted hole on the critical elastic buckling
stress fcr of the isolated web plates and full members with holes from 1106HTable 4.1. These
results are compared to the elastic buckling prediction for a stiffened element with holes
developed and presented in Section 1107H3.3.4. The influence of the hole is minimal for small
hole width to plate width ratios, but increases to a maximum at hhole/h=0.30 for the
ABAQUS plate results (consistent with the stiffened element prediction). fcr increases
with increasing hhole/h for full members, demonstrating that the cross‐section connectivity
decreases a member’s sensitivity to a hole (especially in the range of hhole/h=0.30). The
web is stiffened through beneficial web‐flange interaction created by the relatively stable
edge‐stiffened flange.
As normalized hole width increases, the elastic buckling load exceeds that of a plate
without a hole. This increase in buckling load is attributed to the increased axial stiffness
from the strips adjacent to the hole, which causes local buckling to occur away from the
holes (this “unstiffened strip” effect is discussed in Section 1108H3.3). The length of hole
relative to the length of the member also contributes to the increase in fcr. Lhole/L increases
72
with increasing hhole/h in this study since L=4h. As demonstrated in 1109HFigure 4.3a, the
removal of web material restricts the formation of local buckling in the web of the
member, resulting in shortened half‐waves away from the hole with increased axial
stiffness. The stiffened element prediction in 1110HFigure 4.4 is conservative here when
compared to the plate results because it was developed for evenly spaced holes in long
plates and does not account for the Lhole/L boost in fcr.
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 10
0.2
0.4
0.6
0.8
1
1.2
1.4
1.6
hhole/h
f cr,h
ole/f cr
,no
hole
ABAQUS, SSMA StudABAQUS, stiffened elementStiffened element prediction
Figure 4.4. Effect of a slotted hole on the elastic buckling load of simply supported plates and structural
studs
2.4 70BInfluence of slotted hole location on elastic buckling of an
intermediate length structural stud
This study investigates the elastic buckling behavior of an intermediate length cold‐
formed steel column with one slotted hole. The primary goal here is to identify and
formally define the elastic buckling modes that will be used as predictors of ultimate
strength within the Direct Strength Method. The elastic buckling behavior is compared
73
as the location of a slotted hole is varied along the length of the column. The typical
compression member in this study has a length L of 48 inches and is modeled with an
SSMA 362162‐33 structural channel cross section. A single slotted hole is centered
transversely in the web. The slotted hole has dimensions of hhole=1.5 in., Lhole=4 in., and
rhole=0.75 in. 1111HTable 4.1 summarizes the dimensions of the SSMA 362162‐33 cross section.
The ABAQUS column boundary conditions are consistent with 1112HFigure 4.2.
1113HFigure 4.5 compares the local buckling (L) mode shapes of the column with and
without a slotted hole. The lowest buckling mode is local buckling (L) and exists for
both the column with and without the hole. The location of the hole does not influence
Pcr for this mode as observed in 1114HFigure 4.8. Plate buckling of the web away from the hole
dominates for this mode, regardless of hole location.
Two unique local buckling modes to the column with a hole, LH and LH2, are
also identified in the eigenbuckling analyses. These modes, shown in 1115HFigure 4.5, exhibit
buckling of the unstiffened strip adjacent to the hole similar to that observed in the
cross‐sectional element studies with holes in 1116HChapter 3. The LH mode occurs when both
unstiffened strips buckle in the same direction normal to the web plane, which increases
the distortional tendencies of the flange in the vicinity of the hole. This localized
distortional buckling is observed in 1117HFigure 4.6, which compares the LH modes as the
location of the hole varies along the column length. The LH mode is consistent with the
elastic buckling mode shapes of stiffened elements, where the presence of a hole is
observed to reduce the transverse bending stiffness causing localized deformation at the
hole (see 1118HFigure 3.9a).
74
The LH2 mode occurs when the unstiffened strips buckle in opposite directions
relative to the web plate, resulting in antisymmetric distortional deformation at the hole.
1119HFigure 4.8 demonstrates that Pcr for these two modes is similar and that both modes are
minimally affected by the longitudinal location of the hole in the column.
L LH LH2L
Figure 4.5 The presence of a hole creates unique local buckling modes where unstiffened strip buckling adjacent to the hole occurs symmetrically (LH) or asymmetrically (LH2) increase the distortional tendency
of the flanges
x
C holeL
Figure 4.6 SSMA slotted hole location and local buckling LH mode, L=48 in., x/L=0.06,0.125,0.25,0.375,0.50. Note the distortional tendencies of the flanges at the hole.
75
The pure distortional buckling mode for the column, D, is compared for a column
with and without a hole in 1120HFigure 4.7a. Note that distortional half‐wavelength is
unchanged with the presence of the hole, although local buckling with half‐wavelengths
shorter than the fundamental L half‐wavelength (see 1121HFigure 4.5) mix with the D mode
for the member with the hole. 1122HFigure 4.8 demonstrates that the presence of the hole has
a minimal influence on Pcrd regardless of longitudinal location.
The lowest global buckling mode is flexural‐torsional buckling (GFT) as shown in
1123HFigure 4.7b. The presence of a hole results in a slight decrease in Pcre as the hole moves
towards the end of the column as shown in 1124HFigure 4.8. As the hole shifts close to the
loaded end of the column (x/L=0.06), local buckling at the hole combines with the GFT
mode to reduce Pcre.
D
GFT
Figure 4.7 Influence of a slotted hole on the (a) distortional (D) and (b) global flexural‐torsional (GFT) modes of a cold‐formed steel column
76
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.80
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1
x/L
Pcr
/Py
DGFTLH2LHL
L, no hole
D, no hole
GFT, no hole
Figure 4.8 Influence of SSMA slotted hole location on Pcr for a 362S162‐33 C‐section (refer to 1125HFigure 4.5,
1126HFigure 4.6, and 1127HFigure 4.7 for buckled shape summaries)
2.5 71BFlange hole study
The research focus up until now has been on the elastic buckling modes of
isolated web plates and cold‐formed steel compression members with web holes. Holes
are also commonly present in the flanges of C‐section columns. A standard connection
detail requiring a flange hole is shown in 1128HFigure 4.9, where gypsum sheathing is
connected to steel studs with a bolted or screw attachment (Western States Clay
Products Association 2004).
77
Figure 4.9 Connection detail for structural stud to exterior wall requires a screw or bolt hole placed in the stud flange (Western States Clay Products Association 2004)
This study evaluates the influence of circular flange holes on the elastic buckling
behavior of an intermediate length SSMA 362S162‐33 structural stud. A single hole is
placed at the midlength of both the top and bottom flanges and centered between the
web and lip stiffeners. Five hole diameters consistent with standard bolt holes are
considered: bhole /b =0.178, 0.356, 0.534, 0.713, and 0.892 (¼”,½”,¾”, 1”, 1¼” holes in a 1⅝”
flange) where the flat flange width b=1.40 in. The length L is 48 in. for all members,
corresponding to a common unbraced length of a SSMA structural stud.
1129HFigure 4.10 presents the variation in elastic buckling loads for local, distortional,
and global modes as the diameter of the flange holes increase. The local (L) buckling
load, Pcrl, is not influenced by small holes, but decreases as bhole/b exceeds 0.70. 1130HFigure
4.11 demonstrates that for large flange holes local buckling is dominated by web and
flange deformation near the holes. The large flange holes adversely affect the beneficial
web‐flange interaction inherent in structural studs ( 1131HFigure 4.4 highlights this beneficial
interaction for C‐sections with web holes). The interruption of the web‐flange
interaction by the holes is also reflected in the pure distortional mode (D), as Pcrd
78
decreases slightly as flange hole size increases relative to flange width. The flanges
holes have a minimal effect on the global flexural‐torsional mode (GFT) because their
diameter is small relative to the column length.
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 10
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1
bhole/b
Pcr
/Py
DGFTL
L, no hole
D, no hole
GFT, no hole
Figure 4.10 Influence of flange hole diameter on the local (L), distortional (D), and global (GFT) elastic buckling loads of an SSMA 362S162‐33 structural stud
Figure 4.11 Local (L) buckling is dominated by flange and web deformation near the holes as bhole/b exceeds 0.70
79
4.2.6 72BAnalysis of existing experiments on columns
The Direct Strength Method employs the elastic buckling properties of a cold‐formed
steel member to predict its ultimate strength. To assist in the extension of DSM to
columns with holes, a database is developed in this section which summarizes the elastic
buckling properties and tested strengths of cold‐formed steel columns experiments with
holes from the past 30 years. This database is used 1132HChapter 8 when developing and
verifying DSM for columns with holes. 1133HTable 4.2 summarizes the experimental
programs comprising the database.
Table 4.2 Summary of column experimental data
Author Publication Date Types of Specimens Cross Section End Conditions # of SpecimensOrtiz-Colberg 1981 Stub Column Lipped Cee Channel Fixed-Fixed 8Ortiz-Colberg 1981 Long Column Weak axis pinned 15Miller and Peköz 1994 Stub Column Fixed-Fixed 12Sivakumaran 1987 Stub Column Fixed-Fixed 14Abdel-Rahman 1997 Stub Column Fixed-Fixed 8Pu et al. 1999 Stub Column Fixed-Fixed 9Moen and Schafer 2008 Short Column Fixed-Fixed 6Moen and Schafer 2008 Intermediate Column Fixed-Fixed 6
2.6.1 125BElastic buckling database for column experiments
ABAQUS eigenbuckling analyses were conducted for each specimen in the database.
Member boundary conditions and loading conditions were modeled to be consistent
with the actual experimental conditions. The ABAQUS implementations of the
boundary conditions for each experimental program are described in 1134HFigure 4.12. Local
(L, LH, LH2 – see 1135HFigure 4.5), distortional (D), and global (G) buckling modes were
manually identified from the buckled modes in ABAQUS using the modal definitions
described in 1136H4.2.4.
80
1
2
3
ABAQUS “pinned “ rigid body reference node constrained in 2 to 6 directions, ensures that all nodes on loaded surface translate together in 1 direction but can rotate freely at the platen (no welding)
Nodes bearing on top platen constrained in 1, 2 and 3
45
6
Boundary conditions valid for:
Ortiz-Colberg 1981 (stub columns)Sivakumaran 1987Pu et al. 1999Moen and Schafer 2008
Specimens bear directly on the platens, ends are not welded
ABAQUS “tied “ rigid body reference node constrained in 2 to 6 directions, ensures that all nodes on loaded surface translate together in 1 direction and cannot rotate at the platen (cross-section contact edge is welded)
Nodes bearing on top platen constrained in 1 to 6 (contact edge is welded)
Boundary conditions valid for:
Miller and Pekoz 1994Abdel-Rahman 1997
Specimens are welded to loading platens, cross-section edge rotation is restrained at contact location
ABAQUS “tied “ rigid body reference node constrained in 3 to 6 directions, ensures that all nodes on loaded surface translate together in 1 direction and cannot rotate at the platen (cross-section contact edge is welded). The entire platen can also rotate in the 5 direction.
Boundary conditions valid for:
Ortiz-Colberg 1981 (long column)
Specimens are welded to loading platens, which are attached to pins that allow weak axis rotation of the platens
ABAQUS “tied “ rigid body reference node constrained in 1, 3 to 6 directions, ensures that all nodes on contact surface translate together in 1 direction and cannot rotate at the platen (cross-section contact edge is welded). The entire platen can rotate in the 2 direction.
(a) (b) (c)
Figure 4.12 Experimental program boundary conditions as implemented in ABAQUS 1137HTable 4.3 summarizes the fixed‐fixed column specimen dimensions and material
properties, including cross section and hole dimensions, tested ultimate load Ptest, tested
specimen yield stress Fy, specimen yield force Py,g (calculated with the gross cross‐
sectional area), and Py,net (calculated with the net cross‐sectional area). 1138HTable 4.4
summarizes the fixed‐fixed column specimen elastic buckling loads. ABAQUS
eigenbuckling results are presented for two different types of boundary conditions, the
experiment boundary conditions and CUFSM boundary conditions (warping‐free at the
ends, warping‐fixed at the midlength of the column) except for the Moen and Schafer
specimens which were only modeled with experiment boundary conditions. CUFSM
elastic buckling results are also provided, including the distortional half‐wavelength Lcrd.
The same experiment and elastic buckling information is summarized for the weak‐axis
pinned columns in 1139HTable 4.5 and 1140HTable 4.6 and together with 1141HTable 4.3 and 1142HTable 4.4
comprise the data set used to evaluate the DSM equations for cold‐formed steel columns
with holes 1143HChapter 8.
81
1144HTable 4.7 summarizes cross section and material property parameter ranges for the
fixed‐fixed specimens and weak‐axis pinned specimens. Most of the weak‐axis pinned
specimens are long columns, while the majority of the fixed‐fixed specimens are stub
columns (the exception being the short and intermediate length fixed‐fixed columns
tested by Moen and Schafer). All column specimens in the database are common
industry shapes and meet the DSM prequalification standards (for members without
holes) summarized in 1145HTable 4.8 (AISI‐S100 2007).
82
Table 4.3 Fixed‐fixed column experiment dimensions and material properties
Test Boundary Conditions L t E nu Hole Type L hole h hole r hole H B 1 B 2 D 1 D 2 r F y P y,g P y,net P test
in. in. ksi in. in. in. in. in. in. in. in. in. ksi kips kips kipsOrtiz-Colberg 1981 S4 Fixed-fixed 12.00 0.0492 29420 0.3 Circular 0.75 0.75 0.38 3.50 1.62 1.49 0.49 0.50 0.10 47.1 16.7 14.9 14.2Ortiz-Colberg 1981 S7 12.00 0.0493 29420 0.3 Circular 1.50 1.50 0.75 3.51 1.63 1.49 0.50 0.51 0.10 48.5 17.3 13.7 12.7Ortiz-Colberg 1981 S6 12.00 0.0496 29420 0.3 Circular 1.25 1.25 0.63 3.51 1.61 1.48 0.49 0.51 0.10 51.5 18.4 15.2 13.8Ortiz-Colberg 1981 S8 12.00 0.0496 29420 0.3 Circular 1.75 1.75 0.88 3.51 1.62 1.48 0.49 0.50 0.10 51.5 18.4 13.9 13.6Ortiz-Colberg 1981 S5 12.00 0.0498 29420 0.3 Circular 1.04 1.04 0.52 3.50 1.62 1.48 0.49 0.50 0.10 49.6 17.7 15.2 14.1Ortiz-Colberg 1981 S3 12.00 0.0499 29420 0.3 Circular 0.50 0.50 0.25 3.50 1.61 1.48 0.48 0.50 0.10 49.6 17.7 16.5 14.5Ortiz-Colberg 1981 S14 12.00 0.0760 29420 0.3 Circular 1.04 1.04 0.52 3.52 1.67 1.49 0.51 0.51 0.10 47.4 25.8 22.0 24.6Ortiz-Colberg 1981 S15 12.00 0.0760 29420 0.3 Circular 1.50 1.50 0.75 3.52 1.67 1.49 0.51 0.51 0.10 47.6 25.9 20.4 24.0Abdel-Rahman 1997 A-C 16.73 0.0740 29420 0.3 Circular 2.50 2.50 1.25 7.99 1.64 1.64 0.51 0.51 0.15 55.8 48.3 37.9 26.5*Abdel-Rahman 1997 A-S 16.73 0.0740 29420 0.3 Square 2.50 2.50 --- 7.99 1.64 1.64 0.51 0.51 0.15 55.8 48.3 37.9 26.8*Abdel-Rahman 1997 A-O 18.70 0.0740 29420 0.3 Oval 4.50 2.50 1.25 7.99 1.64 1.64 0.51 0.51 0.15 55.8 48.3 37.9 26.6*Abdel-Rahman 1997 A-R 18.70 0.0740 29420 0.3 Rectangle 4.50 2.50 --- 7.99 1.64 1.64 0.51 0.51 0.15 55.8 48.3 37.9 25.8*Abdel-Rahman 1997 B-C 9.84 0.0500 29420 0.3 Circular 1.50 1.50 0.75 4.00 1.64 1.64 0.51 0.51 0.10 46.2 18.2 14.8 12.7*Abdel-Rahman 1997 B-S 9.84 0.0500 29420 0.3 Square 1.50 1.50 --- 4.00 1.64 1.64 0.51 0.51 0.10 46.2 18.2 14.8 12.7*Abdel-Rahman 1997 B-O 11.81 0.0500 29420 0.3 Oval 4.00 1.50 0.75 4.00 1.64 1.64 0.51 0.51 0.10 46.2 18.2 14.8 12.6*Abdel-Rahman 1997 B-R 11.81 0.0500 29420 0.3 Rectangle 4.00 1.50 --- 4.00 1.64 1.64 0.51 0.51 0.10 46.2 18.2 14.8 12.8*Pu et al. 1999 C-2.0-1-30-1 14.57 0.0787 29420 0.3 Square 1.06 1.06 --- 3.94 2.05 2.05 0.63 0.63 0.16 44.4 30.2 26.5 23.6Pu et al. 1999 C-2.0-1-30-2 14.57 0.0787 29420 0.3 Square 1.05 1.05 --- 3.94 2.05 2.05 0.63 0.63 0.16 33.6 22.8 20.1 18.3Pu et al. 1999 C-2.0-1-30-3 14.57 0.0787 29420 0.3 Square 1.05 1.05 --- 3.94 2.05 2.05 0.63 0.63 0.16 34.4 23.4 20.6 18.3Pu et al. 1999 C-1.2-1-30-1 14.17 0.0472 29420 0.3 Square 1.04 1.04 --- 3.87 2.05 2.05 0.63 0.63 0.11 28.0 11.6 10.3 9.4Pu et al. 1999 C-1.2-1-30-2 14.17 0.0472 29420 0.3 Square 1.04 1.04 --- 3.87 2.05 2.05 0.63 0.63 0.11 28.0 11.6 10.3 9.4Pu et al. 1999 C-1.2-1-30-3 14.17 0.0472 29420 0.3 Square 1.04 1.04 --- 3.87 2.05 2.05 0.63 0.63 0.11 28.0 11.6 10.3 9.4Pu et al. 1999 C-0.8-1-30-1 14.17 0.0315 29420 0.3 Square 1.04 1.04 --- 3.84 2.05 2.05 0.63 0.63 0.08 24.8 7.0 6.2 4.6Pu et al. 1999 C-0.8-1-30-2 14.17 0.0315 29420 0.3 Square 1.04 1.04 --- 3.84 2.05 2.05 0.63 0.63 0.08 24.8 7.0 6.2 4.5Pu et al. 1999 C-0.8-1-30-3 14.17 0.0315 29420 0.3 Square 1.04 1.04 --- 3.84 2.05 2.05 0.63 0.63 0.08 24.8 7.0 6.2 4.6Sivakumaran 1987 A2 7.87 0.0630 29710 0.3 Circular 0.65 0.65 0.32 3.63 1.63 1.63 0.50 0.50 0.13 49.4 22.9 20.9 19.3Sivakumaran 1987 A3 7.87 0.0630 29710 0.3 Square 0.65 0.65 --- 3.63 1.63 1.63 0.50 0.50 0.13 49.4 22.9 20.9 19.0Sivakumaran 1987 A4 7.87 0.0630 29710 0.3 Circular 1.30 1.30 0.65 3.63 1.63 1.63 0.50 0.50 0.13 49.4 22.9 18.8 18.4Sivakumaran 1987 A5 7.87 0.0630 29710 0.3 Square 1.30 1.30 --- 3.63 1.63 1.63 0.50 0.50 0.13 49.4 22.9 18.8 18.3Sivakumaran 1987 A6 7.87 0.0630 29710 0.3 Circular 1.95 1.95 0.97 3.63 1.63 1.63 0.50 0.50 0.13 49.4 22.9 16.8 17.6Sivakumaran 1987 A7 7.87 0.0630 29710 0.3 Square 1.95 1.95 --- 3.63 1.63 1.63 0.50 0.50 0.13 49.4 22.9 16.8 17.4Sivakumaran 1987 A8 8.78 0.0630 29710 0.3 Oval 4.02 1.50 0.75 3.63 1.63 1.63 0.50 0.50 0.13 49.4 22.9 18.2 16.3Sivakumaran 1987 B2 10.43 0.0508 30435 0.3 Circular 1.14 1.14 0.57 6.00 1.63 1.63 0.50 0.50 0.10 38.1 19.0 16.8 12.1Sivakumaran 1987 B3 10.43 0.0508 30435 0.3 Square 1.14 1.14 --- 6.00 1.63 1.63 0.50 0.50 0.10 38.1 19.0 16.8 12.0Sivakumaran 1987 B4 10.43 0.0508 30435 0.3 Circular 2.28 2.28 1.14 6.00 1.63 1.63 0.50 0.50 0.10 38.1 19.0 14.6 12.0Sivakumaran 1987 B5 10.43 0.0508 30435 0.3 Square 2.28 2.28 --- 6.00 1.63 1.63 0.50 0.50 0.10 38.1 19.0 14.6 11.5Sivakumaran 1987 B6 10.43 0.0508 30435 0.3 Circular 3.43 3.43 1.71 6.00 1.63 1.63 0.50 0.50 0.10 38.1 19.0 12.4 10.6Sivakumaran 1987 B7 10.43 0.0508 30435 0.3 Square 3.43 3.43 --- 6.00 1.63 1.63 0.50 0.50 0.10 38.1 19.0 12.4 10.6Sivakumaran 1987 B8 10.43 0.0508 30435 0.3 Oval 4.02 1.50 0.75 6.00 1.63 1.63 0.50 0.50 0.10 38.1 19.0 16.1 11.6Miller & Pekoz 1994 1-12 10.87 0.0756 29420 0.3 Rectangular 2.76 1.61 --- 3.62 1.46 1.46 0.47 0.47 0.09 51.9 27.3 20.9 25.8Miller & Pekoz 1994 1-13 10.87 0.0752 29420 0.3 Rectangular 2.76 1.61 --- 3.62 1.46 1.46 0.47 0.47 0.09 51.9 27.1 20.8 23.6Miller & Pekoz 1994 1-17 17.95 0.0346 29420 0.3 Rectangular 2.24 1.57 --- 5.98 1.34 1.34 0.31 0.31 0.09 44.8 13.9 11.5 5.5Miller & Pekoz 1994 1-19 17.95 0.0346 29420 0.3 Rectangular 2.24 1.57 --- 5.98 1.34 1.34 0.31 0.31 0.09 44.8 13.9 11.5 5.9Miller & Pekoz 1994 2-11 10.87 0.0752 29420 0.3 Rectangular 2.56 1.50 --- 3.62 1.46 1.46 0.47 0.47 0.09 53.0 27.8 21.8 22.2Miller & Pekoz 1994 2-12 10.87 0.0752 29420 0.3 Rectangular 2.56 1.50 --- 3.62 1.46 1.46 0.47 0.47 0.09 53.0 27.8 21.8 22.1Miller & Pekoz 1994 2-14 17.95 0.0350 29420 0.3 Rectangular 2.24 1.57 --- 5.98 1.38 1.38 0.31 0.31 0.09 43.8 13.9 11.5 6.0Miller & Pekoz 1994 2-15 17.95 0.0346 29420 0.3 Rectangular 2.24 1.57 --- 5.98 1.38 1.38 0.31 0.31 0.09 43.8 13.7 11.3 5.8Miller & Pekoz 1994 2-16 17.95 0.0350 29420 0.3 Rectangular 2.24 1.57 --- 5.98 1.38 1.38 0.31 0.31 0.09 43.8 13.9 11.5 5.8Miller & Pekoz 1994 2-24 17.95 0.0354 29420 0.3 Rectangular 2.24 1.57 --- 5.98 1.38 1.38 0.31 0.31 0.09 43.8 14.0 11.6 6.1Miller & Pekoz 1994 2-25 17.95 0.0354 29420 0.3 Rectangular 2.24 1.57 --- 5.98 1.38 1.38 0.31 0.31 0.09 43.8 14.0 11.6 6.1Miller & Pekoz 1994 2-26 17.95 0.0350 29420 0.3 Rectangular 2.24 1.57 --- 5.98 1.38 1.38 0.31 0.31 0.09 43.8 13.9 11.5 6.2Moen and Schafer 2008 362-1-24-H 24.1 0.0391 29500 0.3 Slotted 4.00 1.49 0.75 3.58 1.65 1.60 0.43 0.44 0.20 57.9 16.4 13.0 10.0Moen and Schafer 2008 362-2-24-H 24.1 0.0383 29500 0.3 Slotted 4.00 1.50 0.75 3.64 1.63 1.59 0.44 0.39 0.20 57.1 15.7 12.4 10.4Moen and Schafer 2008 362-3-24-H 24.1 0.0394 29500 0.3 Slotted 4.01 1.49 0.75 3.67 1.67 1.70 0.42 0.43 0.19 56.0 16.4 13.1 9.9Moen and Schafer 2008 362-1-48-H 48.22 0.0393 29500 0.3 Slotted 4.00 1.50 0.75 3.62 1.60 1.60 0.42 0.41 0.19 58.6 16.6 13.1 9.0Moen and Schafer 2008 362-2-48-H 48.23 0.0391 29500 0.3 Slotted 4.00 1.50 0.75 3.62 1.59 1.61 0.42 0.40 0.19 59.7 16.8 13.3 9.2Moen and Schafer 2008 362-3-48-H 48.2 0.0399 29500 0.3 Slotted 4.00 1.49 0.75 3.63 1.60 1.61 0.40 0.43 0.18 58.3 16.8 13.4 9.4Moen and Schafer 2008 600-1-24-H 24.1 0.0421 29500 0.3 Slotted 4.00 1.50 0.75 6.04 1.59 1.61 0.48 0.36 0.16 61.9 25.0 21.1 12.1Moen and Schafer 2008 600-2-24-H 24.1 0.0412 29500 0.3 Slotted 4.00 1.49 0.75 6.01 1.61 1.60 0.37 0.50 0.15 58.4 23.1 19.5 11.6Moen and Schafer 2008 600-3-24-H 24.1 0.043 29500 0.3 Slotted 4.00 1.49 0.75 6.03 1.61 1.58 0.36 0.48 0.16 60.1 24.7 20.9 11.8Moen and Schafer 2008 600-1-48-H 48.09 0.0428 29500 0.3 Slotted 4.00 1.49 0.75 6.01 1.60 1.63 0.48 0.39 0.16 61.4 25.2 21.3 11.2Moen and Schafer 2008 600-2-48-H 48.25 0.0429 29500 0.3 Slotted 4.00 1.50 0.75 6.02 1.59 1.61 0.48 0.36 0.16 62.0 25.5 21.5 11.7Moen and Schafer 2008 600-3-48-H 48.06 0.0431 29500 0.3 Slotted 4.00 1.50 0.75 6.06 1.63 1.59 0.37 0.48 0.16 61.5 25.6 21.6 11.2* average of two tests
Cross Section Dimensions Yield Stress and Force
Study and Specimen Name
Member Material Hole Dimensions
83
Table 4.4 Fixed‐fixed column experiment elastic buckling properties
Boundary Conditions
Pcrl Pcrl, LH Pcrl, LH2 Pcrd Pcre Pcrl Pcrd Pcre Pcrl Pcrd Lcrd
kips kips kips kips kips kips kips kips kips kips in.Ortiz-Colberg 1981 S4 Fixed-fixed 10.7 31.5 --- 40.0 640.0 10.5 17.3 --- 10.8 17.7 13.8Ortiz-Colberg 1981 S7 11.9 25.6 41.9 43.3 640.0 11.1 17.6 --- 10.9 18.0 13.8Ortiz-Colberg 1981 S6 11.6 25.5 33.0 41.8 640.0 11.0 17.7 --- 11.1 18.0 13.8Ortiz-Colberg 1981 S8 12.5 34.6 --- 42.8 640.0 11.4 17.9 --- 11.1 18.0 13.8Ortiz-Colberg 1981 S5 11.4 33.0 41.1 44.5 640.0 10.9 17.8 --- 11.2 18.1 13.8Ortiz-Colberg 1981 S3 11.2 --- --- 41.1 640.0 11.0 18.0 --- 11.3 18.1 13.7Ortiz-Colberg 1981 S14 40.5 45.5 66.9 86.5 964.0 40.1 38.8 --- 39.8 45.5 11.3Ortiz-Colberg 1981 S15 43.5 50.1 --- 70.3 964.0 40.6 40.0 --- 39.8 45.4 11.3Abdel-Rahman 1997 A-C 16.2 12.8 21.6 24.0 1014.2 11.4 14.0 --- 11.7 13.8 15.3Abdel-Rahman 1997 A-S 13.7 22.4 22.5 24.7 1014.2 11.7 14.5 --- 11.7 13.8 15.3Abdel-Rahman 1997 A-O 12.3 16.4 19.5 23.7 811.8 11.1 13.1 --- 11.7 13.8 15.3Abdel-Rahman 1997 A-R 12.9 15.8 19.0 24.1 811.8 11.5 13.1 --- 11.7 13.8 15.3Abdel-Rahman 1997 B-C 11.2 39.5 42.4 45.3 1271.6 9.6 22.3 --- 9.6 16.9 15.9Abdel-Rahman 1997 B-S 12.1 42.0 --- 45.9 1271.6 9.8 23.3 --- 9.6 16.9 15.9Abdel-Rahman 1997 B-O 11.8 23.0 --- 30.6 883.3 9.9 20.8 --- 9.6 16.9 15.9Abdel-Rahman 1997 B-R 12.3 22.7 29.9 30.5 883.3 10.0 20.4 --- 9.6 16.9 15.9Pu et al. 1999 C-2.0-1-30-1 42.6 53.0 85.5 109.0 1182.8 41.1 52.3 --- 42.5 49.9 14.9Pu et al. 1999 C-2.0-1-30-2 42.5 52.9 85.5 109.0 1182.8 41.1 52.3 --- 42.5 49.9 14.9Pu et al. 1999 C-2.0-1-30-3 42.6 52.9 85.5 109.0 1182.8 41.1 52.3 --- 42.5 49.9 14.9Pu et al. 1999 C-1.2-1-30-1 9.5 --- 48.5 50.6 777.4 9.2 19.0 --- 9.4 17.2 22.2Pu et al. 1999 C-1.2-1-30-2 9.5 --- 48.5 50.6 777.4 9.2 19.0 --- 9.4 17.2 22.2Pu et al. 1999 C-1.2-1-30-3 9.5 --- 48.5 50.6 777.4 9.2 19.0 --- 9.4 17.2 22.2Pu et al. 1999 C-0.8-1-30-1 2.8 --- 17.4 17.7 528.3 2.7 11.3 --- 2.8 7.5 27.3Pu et al. 1999 C-0.8-1-30-2 2.8 --- 17.4 17.7 528.3 2.7 11.3 --- 2.8 7.5 27.3Pu et al. 1999 C-0.8-1-30-3 2.8 --- 17.4 17.7 528.3 2.7 11.3 --- 2.8 7.5 27.3Sivakumaran 1987 A2 21.3 --- --- 57.0 2045.6 20.8 35.2 --- 22.0 29.4 13.9Sivakumaran 1987 A3 21.5 --- --- 58.0 2045.6 20.8 35.4 --- 22.0 29.4 13.9Sivakumaran 1987 A4 23.5 --- --- 50.7 2045.6 21.8 38.3 --- 22.0 29.4 13.9Sivakumaran 1987 A5 24.8 --- --- 50.9 2045.6 22.6 39.5 --- 22.0 29.4 13.9Sivakumaran 1987 A6 30.5 --- --- 81.2 2045.6 25.2 35.1 --- 22.0 29.4 13.9Sivakumaran 1987 A7 33.4 --- --- 81.2 2045.6 26.1 36.2 --- 22.0 29.4 13.9Sivakumaran 1987 A8 30.2 30.8 32.2 81.2 1631.1 25.9 33.3 --- 22.0 29.4 13.9Sivakumaran 1987 B2 6.0 --- 10.4 12.0 1742.4 5.8 10.5 --- 5.7 9.8 16.8Sivakumaran 1987 B3 6.1 --- 10.6 12.3 1742.4 5.9 10.6 --- 5.7 9.8 16.8Sivakumaran 1987 B4 7.0 11.2 16.5 19.2 1742.4 6.3 12.5 --- 5.7 9.8 16.8Sivakumaran 1987 B5 7.6 --- 18.6 19.3 1742.4 6.7 13.4 --- 5.7 9.8 16.8Sivakumaran 1987 B6 10.3 20.2 --- 21.1 1742.4 8.1 16.3 --- 5.7 9.8 16.8Sivakumaran 1987 B7 12.4 19.9 --- 21.2 1742.4 9.1 15.8 --- 5.7 9.8 16.8Sivakumaran 1987 B8 6.6 11.7 16.4 18.4 1742.4 5.9 16.5 --- 5.7 9.8 16.8Miller & Pekoz 1994 1-12 43.2 51.1 51.6 76.4 1089.0 37.4 35.5 --- 36.0 42.1 9.7Miller & Pekoz 1994 1-13 42.5 50.3 50.9 75.3 1084.1 36.8 35.0 --- 35.5 41.6 9.7Miller & Pekoz 1994 1-17 1.7 2.4 2.9 3.3 212.7 1.7 2.3 --- 1.7 2.1 8.3Miller & Pekoz 1994 1-19 1.7 2.4 2.9 3.3 212.7 1.7 2.3 --- 1.7 2.1 8.3Miller & Pekoz 1994 2-11 41.3 41.7 47.2 74.4 1084.1 36.5 35.1 --- 35.5 41.6 9.7Miller & Pekoz 1994 2-12 41.3 41.7 47.2 74.4 1084.1 36.5 35.1 --- 35.5 41.6 9.7Miller & Pekoz 1994 2-14 1.8 2.5 3.0 3.4 231.1 1.7 2.4 --- 1.7 2.1 8.3Miller & Pekoz 1994 2-15 1.7 2.4 2.9 3.3 231.1 1.7 2.3 --- 1.7 2.1 8.3Miller & Pekoz 1994 2-16 1.8 2.5 3.0 3.4 231.1 1.7 2.4 --- 1.7 2.1 8.3Miller & Pekoz 1994 2-24 1.9 2.6 3.1 3.5 231.1 1.8 2.4 --- 1.8 2.1 8.3Miller & Pekoz 1994 2-25 1.9 2.6 3.1 3.5 231.1 1.8 2.4 --- 1.8 2.1 8.3Miller & Pekoz 1994 2-26 1.8 2.5 3.0 3.4 231.1 1.7 2.4 --- 1.7 2.1 8.3Moen and Schafer 2008 362-1-24-H 5.9 6.4 --- 9.2 119.3 --- --- --- 5.7 9.5 15.7Moen and Schafer 2008 362-2-24-H 5.4 5.7 --- 10.3 112.8 --- --- --- 5.3 9.0 15.7Moen and Schafer 2008 362-3-24-H 5.7 6.6 --- 9.5 130.6 --- --- --- 5.6 9.3 15.7Moen and Schafer 2008 362-1-48-H 5.3 5.7 --- 9.1 30.0 --- --- --- 5.2 9.0 15.7Moen and Schafer 2008 362-2-48-H 5.2 5.8 --- 9.0 29.7 --- --- --- 5.1 8.9 15.7Moen and Schafer 2008 362-3-48-H 5.7 6.2 --- 9.0 36.2 --- --- --- 5.5 8.9 15.7Moen and Schafer 2008 600-1-24-H 3.3 3.1 --- 7.0 239.3 --- --- --- 3.4 4.9 14.8Moen and Schafer 2008 600-2-24-H 3.2 2.9 --- 6.7 238.4 --- --- --- 3.0 4.9 14.8Moen and Schafer 2008 600-3-24-H 3.5 3.3 --- 7.3 242.6 --- --- --- 3.2 5.0 14.8Moen and Schafer 2008 600-1-48-H 3.4 3.2 --- 5.1 56.3 --- --- --- 3.3 5.1 14.8Moen and Schafer 2008 600-2-48-H 3.4 3.2 --- 5.0 53.0 --- --- --- 3.3 5.1 14.8Moen and Schafer 2008 600-3-48-H 3.4 3.2 --- 5.0 55.8 --- --- --- 3.4 5.1 14.8
Study and Specimen Name
ABAQUS elastic buckling with hole, CUFSM boundary conditions
ABAQUS elastic buckling with hole, experiment boundary conditions CUFSM elastic buckling, no hole
84
Table 4.5 Weak‐axis pinned column experiment dimensions and material properties
Test Boundary Conditions L t E nu Hole Type L hole h hole r hole H B 1 B 2 D 1 D 2 r F y P y,g P y,net P test
in. in. ksi in. in. in. in. in. in. in. in. in. ksi kips kips kipsOrtiz-Colberg 1981 L2 Weak-axis 63.0 0.0490 29420 0.3 Circular 0.50 0.50 0.25 3.51 1.62 1.48 0.50 0.51 0.10 45.7 16.2 15.0 8.5Ortiz-Colberg 1981 L3 pinned 27.0 0.0490 29420 0.3 Circular 1.00 1.00 0.50 3.51 1.62 1.48 0.50 0.51 0.10 42.9 15.2 13.1 11.4Ortiz-Colberg 1981 L6 63.0 0.0490 29420 0.3 Circular 1.00 1.00 0.50 3.51 1.62 1.48 0.50 0.51 0.10 46.1 16.3 14.0 8.5Ortiz-Colberg 1981 L7 63.0 0.0490 29420 0.3 Circular 1.50 1.50 0.75 3.51 1.62 1.48 0.50 0.51 0.10 45.5 16.1 12.7 8.5Ortiz-Colberg 1981 L9 39.0 0.0490 29420 0.3 Circular 1.00 1.00 0.50 3.51 1.62 1.48 0.50 0.51 0.10 43.8 15.5 13.3 9.4Ortiz-Colberg 1981 L10 38.9 0.0490 29420 0.3 Circular 1.50 1.50 0.75 3.51 1.62 1.48 0.50 0.51 0.10 42.3 15.0 11.8 10.1Ortiz-Colberg 1981 L14 39.1 0.0490 29420 0.3 Circular 0.50 0.50 0.25 3.51 1.62 1.48 0.50 0.51 0.10 42.9 15.2 14.1 9.6Ortiz-Colberg 1981 L16 51.0 0.0760 29420 0.3 Circular 1.00 1.00 0.50 3.51 1.62 1.48 0.50 0.51 0.10 48.1 25.9 22.2 17.2Ortiz-Colberg 1981 L17 51.1 0.0760 29420 0.3 Circular 1.50 1.50 0.75 3.51 1.62 1.48 0.50 0.51 0.10 48.1 25.9 20.4 15.0Ortiz-Colberg 1981 L19 27.0 0.0760 29420 0.3 Circular 1.50 1.50 0.75 3.51 1.62 1.48 0.50 0.51 0.10 51.5 27.7 21.9 21.2Ortiz-Colberg 1981 L22 45.0 0.0760 29420 0.3 Circular 1.50 1.50 0.75 3.51 1.62 1.48 0.50 0.51 0.10 46.7 25.1 19.8 20.0Ortiz-Colberg 1981 L26 45.0 0.0760 29420 0.3 Circular 1.00 1.00 0.50 3.51 1.62 1.48 0.50 0.51 0.10 45.8 24.7 21.2 19.1Ortiz-Colberg 1981 L27 27.0 0.0760 29420 0.3 Circular 1.00 1.00 0.50 3.51 1.62 1.48 0.50 0.51 0.10 48.3 26.0 22.3 21.9Ortiz-Colberg 1981 L28 27.0 0.0760 29420 0.3 Circular 1.00 1.00 0.50 3.51 1.62 1.48 0.50 0.51 0.10 42.3 22.8 19.6 22.4Ortiz-Colberg 1981 L32 63.0 0.0760 29420 0.3 Circular 1.00 1.00 0.50 3.51 1.62 1.48 0.50 0.51 0.10 47.9 25.8 22.1 13.3
Study and Specimen Name
Member Material Hole Dimensions Cross Section Dimensions Yield Stress and Force
Table 4.6 Weak‐axis pinned column experiment elastic buckling properties
Boundary Conditions
Pcrl Pcrl, LH Pcrl, LH2 Pcrd Pcre Pcrl Pcrd Pcre Pcrl Pcrd Lcrd
kips kips kips kips kips kips kips kips kips kips in.Ortiz-Colberg 1981 L2 Weak-axis 10.5 16.2 --- 18.7 8.6 10.5 17.9 6.5 10.7 17.7 13.8Ortiz-Colberg 1981 L3 pinned 10.5 13.3 --- 18.7 30.0 10.5 18.9 31.6 10.7 17.7 13.8Ortiz-Colberg 1981 L6 10.5 12.3 --- 18.3 8.6 10.5 17.9 6.5 10.7 17.7 13.8Ortiz-Colberg 1981 L7 10.6 12.5 --- 19.0 8.5 10.5 17.9 6.5 10.7 17.7 13.8Ortiz-Colberg 1981 L9 10.5 14.2 --- 18.6 19.9 10.5 17.7 15.5 10.7 17.7 13.8Ortiz-Colberg 1981 L10 10.7 12.9 16.8 18.6 18.6 10.5 17.2 14.9 10.7 17.7 13.8Ortiz-Colberg 1981 L14 10.5 16.5 16.5 18.5 19.9 10.5 17.7 15.5 10.7 17.7 13.8Ortiz-Colberg 1981 L16 39.6 39.6 --- 46.0 18.5 38.9 44.6 15.4 39.6 45.5 11.3Ortiz-Colberg 1981 L17 40.0 41.5 43.9 46.1 18.2 38.9 44.6 15.3 39.6 45.5 11.3Ortiz-Colberg 1981 L19 41.0 41.4 --- 49.8 50.1 39.0 46.6 45.5 39.6 45.5 11.3Ortiz-Colberg 1981 L22 40.1 42.1 --- 45.0 23.1 38.9 46.1 18.9 39.6 45.5 11.3Ortiz-Colberg 1981 L26 39.7 39.7 --- 45.1 23.5 38.9 46.3 18.9 39.6 45.5 11.3Ortiz-Colberg 1981 L27 39.9 39.9 --- 49.9 52.3 39.0 47.5 45.8 39.6 45.5 11.3Ortiz-Colberg 1981 L28 39.9 43.1 45.8 48.8 56.5 39.0 47.5 45.8 39.6 45.5 11.3Ortiz-Colberg 1981 L32 39.6 39.6 --- 44.5 12.3 38.9 44.5 10.9 39.6 45.5 11.3
ABAQUS elastic buckling with hole, experiment boundary conditions
Study and Specimen Name
CUFSM elastic buckling, no holeABAQUS elastic buckling with hole, CUFSM boundary conditions
85
Table 4.7 Parameter ranges for fixed‐fixed and weak‐axis pinned column specimens with holes Specimen type D/t H/t B/t H/B D/B L/H hhole/h Lhole/L F y (ksi)
min 6.3 46.3 19.3 1.9 0.23 1.7 0.16 0.04 24.8max 20.0 172.7 65.0 4.9 0.32 13.3 0.60 0.46 62.0min 6.6 46.2 20.4 2.3 0.33 7.7 0.16 0.01 42.3max 10.3 71.6 31.7 2.3 0.30 17.9 0.47 0.06 51.5
fixed-fixed columns
weak-axis pinned columns
Table 4.8 DSM prequalification limits for C‐sections
Column parameter DSM prequalification limit
Web slenderness H/t<472Flange slenderness B/t<159 Lip slenderness 4<D/t<33Web / flange 0.7<H/B<5.0Lip / flange 0.05<D/B<0.41Yield stress Fy<86 ksi.
2.6.2 126BBoundary condition influence on elastic buckling
The ABAQUS results in the column elastic buckling database, in addition to serving
as a resource for extending DSM to columns with holes, can also be used to study the
influence of column boundary conditions on elastic buckling. Consider the fixed‐fixed
columns in the database with L/H<4 (most are considered stub columns). 1146HFigure 4.13
and 1147HFigure 4.14 and compare the influence of the experiment fixed‐fixed boundary
conditions for these columns relative to warping free boundary conditions (i.e. CUFSM
style boundary conditions in 1148HFigure 4.2) on Pcrd (distortional buckling) and Pcrl (local
buckling). The experiment boundary conditions are shown to increase Pcrd for all of the
column specimens considered while Pcrl remains relatively unchanged, primarily
because warping deformations are intimately tied to distortional buckling and not plate
buckling (Schafer and Ádàny 2006). For stub columns, the length of the fundamental
distortional half‐wave is often shorter than the length of the column, which results in an
86
increase in Pcrd. The restrained warping at the column ends also contributes to the
shortening of the half‐wave and an increase in Pcrd. The magnitude of this boost in Pcrd
decreases as L/Lcrd increases as shown in 1149HFigure 4.13a because the wavelength shortening
required to accommodate distortional buckling in the column can be distributed over
multiple half‐waves as column length increases. 1150HFigure 4.13b confirms this observation
by demonstrating that Pcrd increases with increasing L/H. H is inversely proportional to
Lcrd for a constant flange width B (i.e., a wider column will have a shorter distortional
half‐wavelength) and therefore as L/H increases, the distortional half‐wavelength
increases relative to the column length causing an increase in Pcrd.
Pcrl increases slightly with increasing hole size relative to column size (both for hhole/h
and Lhole/L) as shown in 1151HFigure 4.14 due to the fixed‐fixed boundary conditions. For large
holes relative to member size the local buckling half‐waves form away from the hole
near the column ends (see Section 1152H3.3). These half‐wavelengths are shortened relative to
their fundamental half‐wavelengths by the loaded column edges which are also
restrained from rotating (from welding), resulting in a higher Pcrl when compared to
warping‐free end conditions with loaded edges free to rotate.
87
0 0.5 1 1.5 2 2.5 30
0.5
1
1.5
2
2.5
3
L/Lcrd
Pcr
d, A
BAQ
US
expe
rimen
t/Pcr
d,AB
AQU
S w
arpi
ng fr
ee
0 0.5 1 1.5 2 2.5 3 3.5 40
0.5
1
1.5
2
2.5
3
L/H
Pcr
d, A
BAQ
US
expe
rimen
t/Pcr
d,AB
AQU
S w
arpi
ng fr
ee
Figure 4.13 Influence of fixed‐fixed boundary conditions versus warping free boundary conditions on Pcrd for column experiments(L/H<4 ) as a function of (a) column length to fundamental distortional half‐
wavelength calculated with CUFSM and (b) column length to member length.
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 10
0.5
1
1.5
2
2.5
3
hhole/h
Pcr
l ,ABA
QU
S ex
perim
ent/P
crl ,A
BAQ
US
war
ping
free
0 0.05 0.1 0.15 0.2 0.25 0.3 0.35 0.4 0.45 0.50
0.5
1
1.5
2
2.5
3
Lhole/L
Pcr
l ,ABA
QU
S ex
perim
ent/P
crl ,A
BAQ
US
war
ping
free
Figure 4.14 Influence of fixed‐fixed boundary conditions versus warping free boundary conditions on Pcrl for column experiments ( L/H<4) as a function of (a) hole width relative to column width and (b) hole length
relative to column length The weak‐axis pinned boundary conditions have a minimal influence on Pcrl and Pcrd
in 1153HFigure 4.15 when compared to the warping‐free boundary conditions. These columns
are still warping‐fixed even though they are pinned (a plate is welded to the end of the
member preventing warping deformation), but because the columns are all relatively
long compared to the stub columns and hole size is small relative to column size, the
wavelength shortening boost in Pcrl and Pcrd is not pronounced.
88
0 0.05 0.1 0.15 0.2 0.25 0.3 0.35 0.4 0.45 0.5
0
0.5
1
1.5
2
2.5
3
Lhole/L
Pcr
l ,ABA
QU
S ex
perim
ent/P
crl ,A
BAQ
US
war
ping
free
0 1 2 3 4 5 60
0.5
1
1.5
2
2.5
3
L/Lcrd
Pcr
d, A
BAQ
US
expe
rimen
t/Pcr
d,AB
AQU
S w
arpi
ng fr
ee
Figure 4.15 Influence of weak‐axis pinned boundary conditions versus warping free boundary conditions on (a) Pcrl as a function of hole length to column length and (b) Pcrd as a function of column length to member
length.
2.7 73BApproximate prediction methods for use in design
The ability to approximate local, distortional, and global critical elastic buckling
loads is central to the extension of the Direct Strength Method (DSM) for cold‐formed
steel structural members with holes. To facilitate the use of DSM for members with
holes, approximate (and conservative) methods for calculating elastic buckling are
developed here which can be used in lieu of a full finite element eigenbuckling analysis.
Elastic buckling approximations using the finite strip method (e.g. CUFSM) are derived
for local and distortional buckling, and modifications to the classical column stability
equations are proposed for global buckling. The simplified methods are intended to be
general enough to accommodate the range of hole shapes, sizes, and spacings common
in industry.
89
2.7.1 127B Local buckling
An approximate method for predicting the local elastic buckling behavior of cold‐
formed steel members with holes is presented in this section. This method extends the
assumption in the “element‐based” methods in 1154HChapter 3 that local buckling occurs as
either plate buckling of the entire cross‐section or unstiffened strip buckling at the
location of the hole. In this finite strip approximate method, local buckling is assumed
to occurs as the minimum of Pcrl occurring from local buckling on the gross cross‐section
(as calculated in the Direct Strength Method) and local buckling of the unstiffened strip
adjacent to the hole. The use of the finite strip method allows for a more realistic
prediction of Pcrl for unstiffened strip buckling by including the interaction of the cross‐
section on the unstiffened strip (i.e., the LH mode for the C‐section in 1155HFigure 4.5). The
method is presented through three examples considering industry standard cross‐
sections with holes which are then verified with ABAQUS thin shell finite element
eigenbuckling results. The prediction method is also validated using the column elastic
buckling database developed in Section 1156H4.2.6.1.
4.2.7.1.1 186BPrediction method
The local critical elastic buckling load Pcrl is calculated for a cold‐formed steel
member with holes as
),min( crhcrcr PPP =l . (4.1)
The calculation of the local critical elastic buckling load on the gross cross‐section, Pcr, is
performed using standard procedures defined in Appendix 1 of the AISI‐S100 (AISI‐
90
S100 2007). Pcrh is calculated in CUFSM using the net cross‐section, which is restrained to
isolate local buckling from distortional buckling by fixing the column cross‐section
corners as shown in 1157HFigure 4.16. It is important to avoid fully restraining a cross‐section
element (i.e., flange or web), since this prevents Poisson‐type deformations and
artificially stiffens the cross‐section. For example, 1158HFigure 4.16a restrains the corners in
the z‐direction only to prevent distortional buckling while still accommodating
transverse deformation of the flanges. The only time a corner should be fixed in both
the x and z directions is when two isolated elements intersect (i.e., C‐section with a
flange hole, see 1159HFigure 4.16a). Finally, when a hole isolates a strip of the net cross‐
section as shown in 1160HFigure 4.16b (e.g., a hat section with flange holes), the isolated
portion of the cross‐section should be deleted since it is assumed to no longer
contributes to the stiffness of the cross‐section over the length of the hole. If the isolated
elements are not removed then the critical elastic buckling load calculated in CUFSM
will correspond to Euler buckling of this isolated portion of the cross‐section.
91
Web hole Flange hole
Web hole Flange hole
Remove isolated elements
ab
Roller (typ.) Pinned (typ.)
Z
X
Figure 4.16 Rules for modeling a column net cross‐section in CUFSM
Once the net‐cross section is restrained, an eigenbuckling analysis is performed, and an
elastic buckling curve similar to 1161HFigure 4.17 is generated. Lcrh is identified on the curve as
the half‐wavelength corresponding to the minimum buckling load. When Lhole<Lcrh as
shown in 1162HFigure 4.17a, Pcrh is equal to the buckling load at the length of the hole. If
Lhole≥Lcrh as shown in 1163HFigure 4.17b, Pcrh is the minimum on the buckling curve. When no
local minimum exists, then Pcrh is equal to the elastic buckling load corresponding to Lhole.
Determining elastic buckling loads at specific half‐wavelengths is a new and
fundamentally different use of the finite strip method when compared to its primary
application within DSM, which is calculating the lowest fundamental elastic buckling
modes of cold‐formed steel members.
92
100 1010
2
4
6
8
10
12
14
16
18
20
half-wavelength
Pcr
100 1010
2
4
6
8
10
12
14
16
18
20
half-wavelength
Pcr
half-wavelength, inches
P cr,
kips
Lhole<Lcrh
Pcrh
Lcrh
half-wavelength, inches
Lhole>Lcrh
Pcrh
Lcrh
Figure 4.17 Local elastic buckling curve of net cross‐section when (a) hole length is less than Lcrh and (b) when hole length is greater than Lcrh
4.2.7.1.2 187BMethod examples
Three examples are presented here that approximate the local critical elastic buckling
load Pcrl for cold‐formed steel columns with holes using CUFSM. For all examples, the
length of the column L=100 inches and five slotted holes are spaced at S=20 inches. The
typical length of the hole Lhole=4 inches. All ABAQUS eigenbuckling analyses are
modeled with CUFSM‐style boundary and loading conditions identical to those shown
in 1164HFigure 4.2. The modulus of elasticity, E, is assumed as 29500 ksi and Poisson’s ratio,
ν, as 0.3 in all finite strip and finite element models. Pcrl is normalized when plotted by
Py,g, the squash load of the column calculated with the gross cross‐sectional area and a
yield stress, Fy, of 50ksi.
The first example is an SSMA 362S162‐33 cross section with a slotted web hole.
1165HFigure 4.18 compares the finite strip and ABAQUS mode shapes for hhole/hC=0.14 where hC
is the C‐section web depth measured from the centerline flange to centerline flange. The
CUFSM approximate method predictions are plotted for a range of hhole/hC in 1166HFigure 4.19,
93
and compared with ABAQUS eigenbuckling predictions to demonstrate the viability of
the prediction method. For this example, smaller hole widths lead to reductions in Pcrl
when compared to a member without a hole or members with larger holes. This
counterintuitive result occurs because for small holes the unstiffened strip controls the
local buckling behavior (i.e., the LH mode) and for large holes, local plate buckling
occurs between the holes (i.e., the L mode), which is consistent with the elastic buckling
observations for plates (see 1167HChapter 3). (One must keep in mind that for strength the net
section in yielding, as well as the elastic buckling load, ultimately determine the
capacity, not just Pcrl.)
CUFSM Approximation(SSMA 362S162-33)
ABAQUS
Lcrh
hhole hC
Figure 4.18 Comparison of CUFSM and ABAQUS predictions of unstiffened strip buckling.
94
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 10
0.05
0.1
0.15
0.2
0.25
0.3
0.35
0.4
0.45
0.5
hhole/hC
Pcr
l /P
y,g
ABAQUSCUFSM Approx. Method
Figure 4.19 ABAQUS results verify CUFSM local buckling predictions for an SSMA 362S162‐33 column
with evenly spaced web holes.
The next example evaluates the influence of a slotted flange hole on Pcrl for an SSMA
362S162‐33 cross section. The unstiffened strip buckled mode shape for this cross‐
section from both finite strip and finite element predictions are compared in 1168HFigure 4.20.
It is observed that for both CUFSM and ABAQUS mode shapes, buckling occurs
primarily in the web and flange strip, and that the flange strip – lip portion of the cross‐
section remains stable at Pcrh. The CUFSM prediction method results are plotted for
varying flange hole width bhole relative to centerline flange width bC and compared to
ABAQUS eigenbuckling predictions in 1169HFigure 4.21. Pcrh decreases with increasing flange
hole width for both CUFSM and ABAQUS results. The decreasing trend in the critical
elastic buckling load demonstrates the importance of the flange in web local buckling
dominated cross‐sections.
95
CUFSM Approximation(362S162-33)
ABAQUS
Lcrl
bhole
Assume Lcrl=Lholein CUFSM
bC
Figure 4.20 CUFSM and ABAQUS local buckling mode shapes are consistent when considering a slotted
flange hole.
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 10
0.05
0.1
0.15
0.2
0.25
0.3
0.35
0.4
0.45
0.5
bhole/bC
Pcr
l /P
y,g
ABAQUSCUFSM Approx. Method, Lcrl=Lhole
Figure 4.21 ABAQUS results verify CUFSM predictions for an SSMA 362S162‐33 cross section with evenly
spaced flange holes. The third example is an SSMA 1200S162‐68 cross section with a slotted hole centered
in the web. 1170HFigure 4.22 provides the CUFSM and ABAQUS buckled shapes when
hhole/hC=0.16. The assumption in the CUFSM prediction method that Lcrl is equal to Lhole=4
in. produces a Pcrh higher than Pcr without the hole (because Lcrl is shorter than the local
96
buckling half‐wavelength of the column) and therefore Pcr controls in the prediction
method as shown in 1171HFigure 4.23. The approximate method correctly predicts that
unstiffened strip buckling does not occur as observed in the ABAQUS buckled shape,
and that the actual local buckling half‐wavelength Lcrl is similar to that of a column
without holes. The prediction for Pcrl is unconservative here though (ABAQUS results
are 10% lower than Pcr), because the hole causes a mixed local‐distortional mode that is
not captured by the CUFSM net‐section model (with pinned corners) or the CUFSM
gross cross‐section model (without the influence of the web hole). For sections such as
this where local and distortional buckling have similar half‐wavelengths and critical
elastic buckling loads, a full finite element eigenbuckling analysis may be warranted to
evaluate the presence of holes.
CUFSM Approximation(SSMA 1200S162-68)
ABAQUS
hhole
Distortional buckling mixes with local buckling
Figure 4.22 ABAQUS predicts local plate buckling with distortional buckling interaction which is not
detected in CUFSM.
97
0 0.1 0.2 0.3 0.4 0.50
0.05
0.1
0.15
0.2
0.25
0.3
0.35
0.4
0.45
0.5
hhole/hC
Pcr
l /P
y,g
ABAQUSCUFSM Approx. Method Lcrl=Lhole
Pcrh from CUFSM is always higher than Pcr for this cross-section
Pcr
Figure 4.23 ABAQUS results are slightly lower than CUFSM predictions, CUFSM predicts correctly that
plate local buckling controls over unstiffened strip buckling.
4.2.7.1.3 188BMethod validation using elastic buckling database
The elastic buckling database developed in Section 1172H4.2.6.1 is utilized here to evaluate
the CUFSM approximate method for predicting Pcrl. Custom Matlab code was written to
calculate Pcrh for all 78 specimens in the database (Mathworks 2007). The code performed
a CUFSM analysis of the net cross‐section (cross‐section containing a hole) with pinned
corners (x‐ and z‐directions). The predicted Pcrl of each column specimen is the
minimum of Pcrh (unstiffened strip buckling at the net section) and Pcr ( 1173HTable 4.4 and
1174HTable 4.6, CUFSM elastic buckling results, no hole).
1175HFigure 4.24 compares Pcrl reflecting the experimental boundary conditions in
ABAQUS (from 1176HTable 4.4 and 1177HTable 4.6) relative to Pcr and Pcrh. For all specimens, Pcr (no
hole, gross cross‐section) is lower than Pcrh (hole, net cross‐section) because the strips of
web adjacent to the hole are stiffer than the cross‐section away from the holes (similar to
98
the SSMA362S162‐33 cross‐section with hhole/h>0.20, see 1178HFigure 4.19). Even for those
column specimens with small holes relative web width, the holes are often circular or
square and therefore Pcrh is predicted higher than Pcr since the buckling half‐wavelength
of the unstiffened strip is assumed equal to the diameter of the hole. This prediction
result is consistent with the actual buckled behavior of stiffened elements with circular
and square circular holes shown in 1179HFigure 3.19.
1180HFigure 4.25 compares the ABAQUS experiment Pcrl to the predicted Pcrl and
demonstrates the approximate method is accurate for smaller holes relative to column
size and becomes increasing conservative as hole size increases relative to column size
(hhole/h and Lhole/L). The prediction becomes conservative because it does not take into
account the wavelength stiffening effects (discussed in Section 1181H3.3.2) which boost Pcrl as
the hole becomes large relative to the column. The mean and standard deviation of the
ABAQUS to predicted ratio are 1.11 and 0.18 respectively, demonstrating the viability of
the method for the specimens considered.
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 10
0.5
1
1.5
2
2.5
3
hhole/h
Pcr
l ,ABA
QU
S/Pcr
l ,pre
dict
ed
Pcrh (CUFSM net section)
Pcr (CUFSM gross section)
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 10
0.5
1
1.5
2
2.5
3
Lhole/L
Pcr
l ,ABA
QU
S/Pcr
l ,pre
dict
ed
Pcrh (CUFSM net section)
Pcr (CUFSM gross section)
Figure 4.24 Predicted Pcrh (CUFSM, buckling of the net cross‐section) and Pcr (CUFSM, buckling of the gross cross section, no hole) are compared relative to the ABAQUS Pcrl with experiment boundary conditions as a
function of (a) hole width to flat web width and (b) hole length to column length
99
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 10
0.5
1
1.5
2
2.5
3
hhole/h
Pcr
l ,ABA
QU
S/Pcr
l ,pre
dict
ed
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 10
0.5
1
1.5
2
2.5
3
Lhole/L
Pcr
l ,ABA
QU
S/Pcr
l ,pre
dict
ed
Figure 4.25 Predicted Pcrl (CUFSM approximate method) is compared relative to the ABAQUS Pcrl with experiment boundary conditions as a function of (a) hole width to flat web width and (b) hole length to
column length
2.7.2 128B Distortional buckling
An approximate method utilizing the finite strip method is developed here for
predicting the distortional critical elastic buckling load Pcrd of cold‐formed steel columns
with holes. The method simulates the loss in bending stiffness of a cross‐section within
a distortional buckling half‐wave by modifying the cross‐section thickness in CUFSM.
Two different approaches to simulating this loss in stiffness are evaluated. The first
approach reduces the member thickness in the regions of the cross‐section with holes
based on the ratio of hole length to distortional half‐wavelength. The second approach
is developed for C‐sections with web holes and is mechanics‐based, where the thickness
of the entire web is reduced based on an assumed relationship between web bending
stiffness (derived with observations from ABAQUS thin shell elastic FE analyses) and
the bending stiffness matrix terms of a finite strip element. The steps for implementing
these methods in CUFSM are described, and an example is provided where the
prediction method for Pcrd is compared to ABAQUS eigenbuckling results for an industry
100
standard SSMA 250S162‐68 cross‐section with evenly spaced web holes along the length
of a column. An empirical equation is derived to account for the increase in Pcrd from
warping fixed end conditions and then the viability of the approximate method is
evaluated against Pcrd from the column experiment database in Section 1182H4.2.6.1.
4.2.7.2.1 189BPrediction method
The prediction method presented here for Pcrd assumes that the change in cross‐
sectional stiffness within a distortional half‐wave caused by the presence of a hole (or
holes) can be simulated by assuming a reduced thickness of the cross‐section. The
distortional half‐wavelength of the cross‐section, Lcrd, without holes is determined first.
The elastic buckling curve is calculated using the gross section of the column in CUFSM
and Lcrd is read off of the curve at the location of the distortional local minimum as
shown in 1183HFigure 4.26 (this elastic buckling curve corresponds to an SSMA 250S162‐68
cross section, where Lcrd=12 in.). The prediction method assumes that Lcrd does not change
with the presence of holes. The cross‐section is then modified to approximate the
presence of holes within a distortional half‐wavelength. Two approaches for this
modification step are presented next in Section 1184H4.2.7.2.1.1 and Section 1185H4.2.7.2.1.2. Once
the cross section is modified to account for the presence of a hole in CUFSM, another
elastic buckling curve is generated and Pcrd (including the presence of the hole) is
determined as the elastic buckling load occurring at Lcrd as shown in 1186HFigure 4.26.
101
100 101 1020
10
20
30
40
50
60
70
80
90
100
half-wavelength, in.
Pcr
, kip
s
without holewith hole
100 101 1020
10
20
30
40
50
60
70
80
90
100
half-wavelength, in.
Pcr
, kip
s
without holewith hole
Lcrd (determined at local minimum of no hole curve)
Pcrd (includes influence of hole)
Figure 4.26 CUFSM approximate method for calculating Pcrd for a column with holes.
.7.2.1.1 209B“Weighted average” approach for predicting hole influence
The hole influence on distortional buckling of an open thin‐walled cross section can
be approximated by modifying the cross‐section thickness in CUFSM at the location of a
hole with the following equation:
tLL
tcrd
holehole ⎟⎟
⎠
⎞⎜⎜⎝
⎛−= 1 . (4.2)
The implementation of the reduced thickness in a C‐section with a single web hole is
provided in 1187HFigure 4.27. This approach is an intuitive first cut at approximating the
reduction in bending stiffness of the cross section. A more refined mechanics‐based
approach is presented next.
102
hhole
t
thole
Figure 4.27 Modified cross section to be used in CUFSM to predict Pcrd for a column with holes.
.7.2.1.2 210B“Mechanics-based” approach for predicting hole influence A plate model is developed in ABAQUS to study the influence of a hole on the
bending stiffness of a SSMA 250S162‐68 column experiencing distortional buckling. The
stiffness reduction observed in ABAQUS is quantified and then equated to finite strip
bending stiffness matrix terms to derive a reduced web thickness expression based on
finite strip mechanics. The plate dimensions in ABAQUS are chosen to correspond to
the web of the 250S162‐68 section over one distortional half‐wave. The plate width h is
2.4 in., the plate length L=12 inches (consistent with Lcrd=12 in.), and t=0.0713 in. One
slotted hole with Lhole=4 in. is centered in the plate. The width of the hole is varied,
hhole=0.5 in., 0.96 in., 1.20 in., 1.5 in., and 1.75 in. (and subsequently rhole varies). The
modulus of elasticity, E, is assumed as 29500 ksi and Poisson’s ratio, ν, as 0.3 for all finite
element models considered here. The ABAQUS boundary conditions and applied
loading are described in 1188HFigure 4.28. The plate is simply‐supported and loaded with
imposed rotations at the long edges of the plate with magnitudes varying as a half‐sine
wave to simulate distortional deformation over one half‐wavelength.
103
1 (x)
2
3
Restrain perimeter in 2
Restrain midline node in 1 and 3
Restrain midline node in 3
Apply imposed rotation at plate edges in the shape of 0.001*sin(πx/L) radians
Figure 4.28 ABAQUS boundary conditions and imposed rotations for web plate The deformed shape of the plate when hhole/h=0.50 is provided in 1189HFigure 4.29. At each
node where an imposed rotation is applied, the associated moment is obtained from
ABAQUS and plotted in 1190HFigure 4.30 as a transverse bending stiffness per unit length.
(Note that near x=0 in. and x=12 in., the deformed shape in ABAQUS results in a small
negative bending stiffness which is not plotted in 1191HFigure 4.30 and does not affect the
overall results here. The negative stiffness is not predicted in the finite strip formulation
because the longitudinal shape function is enforced as a half‐sine wave). The hole
causes a sharp reduction in bending stiffness at the location of the hole, but has a
minimal influence on bending stiffness away from the hole. The stiffness reduction is
shown to be relatively insensitive to the ratio of hole width to plate width except for
peaks in stiffness that increase with hhole/h at the rounded edges of the slotted hole. The
results in 1192HFigure 4.30 confirm the intuitive assumption employed to develop Eq. 1193H(4.2);
the ratio of the length of the hole to the length of the distortional half‐wave is an
important parameter when predicting the loss in bending stiffness.
104
Figure 4.29 Plate deformation from imposed edge rotations, hhole/h=0.50
0 2 4 6 8 10 120
0.01
0.02
0.03
0.04
0.05
0.06
0.07
distance along plate, in.
rota
tiona
l stif
fnes
s, k
*in/ra
d
no hole
Increasing hhole/h
hhole/h=0.63
Figure 4.30 Transverse rotational stiffness of the plate is significantly reduced in the vicinity of the slotted hole
If K represents the cumulative transverse bending stiffness for the plate without a
hole (area under the curve in 1194HFigure 4.30), then the reduced K including the presence of
the hole can be approximated as:
KLLK
crd
holehole ⎟⎟
⎠
⎞⎜⎜⎝
⎛−= 1 . (4.3)
105
The global bending stiffness K for a simply‐supported finite strip element is derived by
applying a unit rotation at the strip edges:
[ ]⎪⎪⎭
⎪⎪⎬
⎫
⎪⎪⎩
⎪⎪⎨
⎧
=
⎪⎪⎭
⎪⎪⎬
⎫
⎪⎪⎩
⎪⎪⎨
⎧
−⎥⎥⎥⎥
⎦
⎤
⎢⎢⎢⎢
⎣
⎡
=
KVKV
kkkkkkkkkkkkkkkk
dk Tweb
1010
44434241
34333231
24232212
14131211
θ , (4.4)
where the keb is the bending stiffness matrix and dwθ=[w1 θ1 w2 θ2] (Schafer and Ádàny
2006). Solving Eq. 1195H(4.4) for K:
2422 kkK −= . (4.5)
Since k22 and k24 are both functions of the web thickness (tweb)3 , K and Khole can be equated
directly as:
( )
( )33
,
web
holewebhole
t
tK
K= . (4.6)
Substituting Eq. 1196H(4.6) and rearranging in terms of tweb,hole, the reduced web thickness
corresponding to the reduced transverse bending stiffness from the hole is:
webcrd
holeholeweb t
LL
t3/1
, 1 ⎟⎟⎠
⎞⎜⎜⎝
⎛−= . (4.7)
Eq. 1197H(4.7) is an improvement over Eq. 1198H(4.2) because it reflects the underlying plate
bending mechanics involved in distortional buckling and is actually simpler to
implement in CUFSM since the entire web thickness of a C‐section is reduced to tweb,hole
instead of changing the sheet thickness just at the location of the hole as shown in 1199HFigure
4.27. A similar modification to t has been proposed for web‐slotted thermal structural
studs (Kesti 2000).
106
4.2.7.2.2 190BMethod example
The distortional critical elastic buckling load Pcrd is calculated here with the CUFSM
prediction method for a long column (L=100 in.) with an SSMA 250S162‐68 cross‐section
and five evenly spaced slotted web holes (S=20 in., Lhole=4 in.). The width of the hole is
varied relative to the web width, and ABAQUS eigenbuckling results are used to
evaluate the viability of the method. All ABAQUS finite element models have CUFSM
style boundary and loading conditions as shown in 1200HFigure 4.2. The modulus of
elasticity, E, is assumed as 29500 ksi and Poisson’s ratio, ν, as 0.3 in all finite strip and
finite element models. Pcrd is normalized by Py,g when plotted. Py,g is the squash load of
the column calculated with the gross cross‐sectional area and assuming Fy=50 ksi.
A comparison of the CUFSM prediction method (employing the “weighted average”
thickness approximation) and ABAQUS distortional buckling mode shapes are provided
in 1201HFigure 4.31 when hhole/h=0.63. Nine distortional half‐waves form along the member in
ABAQUS, with every other half‐wave containing one slotted hole. The CUFSM
prediction method employing both the “weighted average” and “mechanics‐based”
thickness modifications to the cross‐section are compared over a range of hhole/h to
ABAQUS eigenbuckling results in 1202HFigure 4.32. Pcrd for the pure ABAQUS distortional
(D) buckling mode is plotted to demonstrate that prediction method is viable for this
cross‐section and hole spacing. The CUFSM prediction for Pcrd with the “weighted
average” thickness reduction at the hole decreases with increasing hole width since the
web provides less bending stiffness to the flanges as more hole material is removed. The
CUFSM prediction employing the “mechanics‐based” reduction in web thickness is not
107
a function of hhole/h and is shown to be a more realistic predictor of Pcrd than the “weighted
average” approach. These approximate methods are evaluated against ABAQUS
distortional buckling predictions from the column database in Section 1203H4.2.7.2.4.
CUFSM Approximation(250S162-68)
ABAQUS
hhole
Lcrd
Figure 4.31 Comparison of CUFSM and ABAQUS distortional buckling mode shapes.
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 10
0.5
1
1.5
2
2.5
3
hhole/h
Pcr
d/Py,
g
CUFSM Approx. Method with "weighted average" thole
CUFSM Approx. Method with "mechanics-based" tw eb,hole
ABAQUS D mode
Figure 4.32 CUFSM distortional buckling prediction method is conservative when considering an SSMA
262S162‐68 column with uniformly spaced holes.
108
4.2.7.2.3 191BWarping‐fixed distortional amplification factor
Longitudinal warping deformations occur as a result of distortional buckling in cold‐
formed steel columns. When this warping deformation is restrained, the distortional
buckling half‐wavelength is shortened relative to the fundamental distortional half‐
wavelength of the column cross‐section, Lcrd. This change in half‐wavelength results in
an amplification of the distortional critical elastic buckling load of the column, Pcrd (see
1204HFigure 4.13a for boost in Pcrd for stub columns). The elastic buckling database developed
in Section 1205H4.2.6.1 (1206HTable 4.4 and 1207HTable 4.6) provides an opportunity to derive an
empirical amplification factor since for all columns in the database, Pcrd for both warping‐
fixed (ABAQUS) and warping free (CUFSM) boundary conditions are known and the
fundamental distortional half‐wavelength, Lcrd, has been calculated in CUFSM.
The warping‐fixed boundary condition effect on Pcrd is plotted for the 78 specimens in
the column database in 1208HFigure 4.33. The boost in Pcrd is highest when the column is short
relative to Lcrd because the wavelength shortening must be accommodated over just one
half‐wave. An empirical equation (also plotted in 1209HFigure 4.33) is fit to the data trend:
2
boost 211D ⎟
⎠⎞
⎜⎝⎛+=
LLcrd (4.8)
This amplification factor can be used with the CUFSM prediction method developed in
Section 1210H4.2.7.2.1 when the column being considered has warping‐fixed boundary
conditions. Eq. 1211H(4.8) is consistent with the distortional buckling boost factor provided in
the DSM Design Guide (AISI 2006) as shown in 1212HFigure 4.33.
109
0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 50.5
1
1.5
2
2.5
3
L/Lcrd
Pcr
d,fix
ed-fi
xed
/Pcr
d,C
UFS
M w
arpi
ng fr
ee
Specimen dataEq.(4.8)DSM Design Guide
Figure 4.33 Warping‐fixed boundary condition amplification of Pcrd
4.2.7.2.4 192BVerification of CUFSM approximate method with column database
The CUFSM approximate method for distortional buckling is now evaluated using
the elastic buckling properties of the 78 column specimens from the experiment database
developed in 1213H4.2.6.1. The ABAQUS distortional critical elastic buckling load Pcrd,
determined with the experiment boundary conditions, is plotted against the
approximate method predictions in 1214HFigure 4.34. The predictions including the
distortional amplification factor from Eq. 1215H(4.8). The approximate method employing the
“weighted average” reduced web thickness at the hole from Eq. 1216H(4.2) and the
“mechanics‐based” reduced web thickness approach from Eq. 1217H(4.7) are both presented.
The accuracy of the prediction method improves as the column length increases relative
to the fundamental distortional half‐wavelength Lcrd. The prediction accuracy is highly
110
variable when L/Lcrd<1, primarily because of the variation in the boundary condition
influence described in Section 1218H4.2.7.2.3 for stocky columns. As expected, the
“mechanics‐based” thickness approach (with ABAQUS to predicted ratio mean and
standard deviation of 1.19 and 0.29) is more accurate over the 78 columns than the
“weighted average” approach (mean of 1.24 and standard deviation of 0.29).
0 1 2 3 4 5 60
0.5
1
1.5
2
2.5
3
L/Lcrd
Pcr
d,AB
AQU
S Ex
p. B
C/Pcr
d,pr
edic
ted
weighted average thole
mechanics-based tw eb,hole
Figure 4.34 Accuracy of the CUFSM approximate method for predicting Pcrd improves as column length increases relative to the fundamental distortional half‐wavelength for warping‐fixed columns
2.7.3 129B Global buckling
Two different approximate methods for calculating the critical elastic buckling load
of a column, Pcre, are evaluated in this section. Both methods employ weighted averages
of the member section properties in the classical column stability equation to account for
the influence of holes, one using a weighted cross‐sectional thickness at the locations of
the hole and the other using the weighted average of the gross and net cross sections. The
111
approximate methods are compared to ABAQUS eigenbuckling results for a long cold‐
formed steel column (SSMA 1200S162‐68 cross section) with uniformly spaced circular
holes. The average torsional constants, J and Cw, are calculated directly using ABAQUS
for this column and then compared to their associated weighted average estimates.
Based on these studies, recommendations are made regarding the approximate method
most suitable for predicting Pcre for columns with holes.
4.2.7.3.1 193BDescription of methods
.7.3.1.1 211BWeighted Properties Method The equation for predicting the global (flexural only) critical elastic buckling load Pcre
of a column with holes along its length can be solved using energy methods, and is
derived for a column with two holes located symmetrically about the longitudinal
midline of the column in 1219HAppendix E. The equation that evolves from the Raleigh‐Ritz
derivation for this case is:
⎟⎟⎠
⎞⎜⎜⎝
⎛ +=
LLILI
LEP HnetNHg
cre 2
2π (4.9)
where Ig is the moment of inertia of the gross cross‐section, Inet is the moment of inertia of
the net cross‐section, LNH is the length of column without holes and LH is the length of
column with holes (note that LNH + LH= L). The average moment of inertia of the column
with holes is shown in Eq. 1220H(4.9) to be equivalent to the weighted average of the gross
and net cross sections along the column length. An approximate method for calculating
Pcre is proposed here which extends this “weighted properties” methodology in Eq. 1221H(4.9)
112
to all of the cross‐section properties of the column required to solve the classical cubic
buckling equation for columns (Chajes 1974):
( )( )( ) ( ) ( ) 02
22
,2
22
,,,, =−−−−−−−o
oxcre
o
oycrecrexcreycre r
yPPP
rxP
PPPPPPPP φ (4.10)
including the cross‐sectional area A, moment of inertia Ix and Iy, St. Venant torsional
constant J, and shear center location. The computer program CUTWP solves Eq. 1222H(4.10)
for any general cross‐section and is freely available (Sarawit 2006). The net section
properties can be calculated in CUFSM (or CUTWP) by reducing the sheet strip
thickness to zero at the location of the hole. This approximation is referred to as the
“weighted properties” method. A general form of Eq. 1223H(4.9) is also derived in 1224HAppendix
E which can be used with the “weighted properties” method for the case of a column
with a single hole or multiple arbitrarily‐spaced holes.
.7.3.1.2 212BWeighted Thickness Method This approximate method approaches the calculation of the average column section
properties in a different way, by using a weighted member thickness at the location of
the holes in the cross‐section to calculate the average cross sectional properties:
tLLLt H
hole−
= . (4.11)
An example of a C‐section with a reduced thickness at the location of a web hole is
provided in 1225HFigure 4.35. This “weighted thickness” method is more convenient to
implement than the “weighted properties” method presented in the previous section
113
because the modified cross‐section (with reduced thickness) can be input directly into a
computer program such as CUTWP.
hhole
t
thole
Figure 4.35 A “weighted thickness” cross section can be input directly into a program that solves the classical cubic stability equation for columns (e.g. CUTWP).
4.2.7.3.2 194BExample and verification
ABAQUS global eigenbuckling results are compared in this section to the “weighted
properties” and “weighted thickness” prediction methods for an industry standard
SSMA 1200S162‐68 long column (SSMA 2001) with evenly spaced circular holes. The
length of the column L=100 in., the hole spacing S=20 in., and the diameter of the circular
hole is varied from hhole/H=0.10 to hhole/H=0.90 where H is the out‐to‐out depth of the cross‐
section (see 1226HFigure 4.1 for cross‐section dimension notation). All ABAQUS finite
element models are loaded in compression at the member ends and have warping free
boundary conditions as shown in 1227HFigure 4.2. The modulus of elasticity, E, is assumed as
29500 ksi and Poisson’s ratio, ν, as 0.3 in all CUTWP and finite element models.
The three global buckling modes of this SSMA 1200S162‐68 long column without
holes are calculated in CUTWP: (1) weak axis flexural buckling occurs at Pcr=7.91 kips,
(2) flexural‐torsional buckling occurs at Pcr=13.39 kips, and (3) strong‐axis flexural
114
buckling occurs at Pcr=604.17 kips. The first two buckling modes are the focus of this
study since the strong‐axis buckling mode is much higher than the squash load of the
column (Py,g= 56.30 kips assuming Fy=50 ksi) and will not influence the design of the
column. 1228HFigure 4.36 provides an example of the weak‐axis flexural and flexural‐
torsional buckling modes when hhole/H=0.50. Note that shell FE predicts local buckling
mixing with the weak‐axis flexural mode when hhole/H>0.50 because Pcre is reduced by the
presence of holes to a magnitude similar to the local critical elastic buckling load
(Pcrl=6.69kips). Local buckling is not observed to mix with global buckling in the
flexural‐torsional (column) or lateral‐torsional (beam) buckling in this study.
Global column buckling modes (SSMA 1200S162-68, hhole/h=0.50)
Weak Axis FlexuralPcre=6.96 kips
Flexural-TorsionalPcre=10.64 kips
Web local buckling mixes with global mode
Figure 4.36 Weak‐axis flexural and flexural‐torsional global buckling modes for an SSMA 1200S162‐68
column with evenly spaced circular holes.
.7.3.2.1 213BSection property calculations at the net section To draw meaningful conclusions regarding the “weighted properties” and
“weighted thickness” prediction methods it is first helpful to understand how circular
hole diameter influences the column’s section properties. 1229HFigure 4.37 compares the net
section to the gross cross‐sectional area A, the strong and weak axis moment of inertia Ix
and Iy, the St. Venant torsional constant J, and the warping torsional constant Cw of the
115
column as they vary with hhole/H. All net section properties in this figure are determined
with the CUFSM section property calculator by reducing the sheet thickness to zero at
the location of the hole. A and J decrease linearly with hole diameter while Ix and Iy
decrease nonlinearly. Iy is most sensitive to the presence of the hole because the hole is
located in the channel web for this case, which provides much of the contribution to the
weak axis moment of inertia. Cw, calculated here assuming zero thickness but continuity
through the hole, varies nonlinearly with hhole/H. It is unclear if the net section Cw
calculated in this way produces the best approximation of the column’s actual physical
behavior in torsional buckling. The magnitude of Cw is influenced heavily by cross‐
section continuity since a line integral around the cross‐section is used to solve for the
warping function. Further investigation of J and Cw for columns with holes is presented
in Section 1230H4.2.7.3.2.3.
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 10
0.5
1
1.5
hhole/H
net s
ectio
n/gr
oss
sect
ion
Cw
Ix (strong)
Iy (weak)
AJ
Figure 4.37 Variation in net section properties as circular hole diameter increases.
116
.7.3.2.2 214BAverage section property calculations for the column - A, Ix, and Iy
The average section properties of the 1200S162‐68 column with circular holes
calculated using the “weighted thickness” and “weighted properties” methods are
compared in 1231HFigure 4.38 through 1232HFigure 4.40. For this example problem there are
minimal differences between the methods for A and Ix, although Iy calculated with the
“weighted properties” method decreases in stiffness relative to the “weighted thickness”
method as hole diameter increases relative to column width.
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 10
0.5
1
1.5
hhole/H
Aav
g/Ag
weighted thicknessweighted properties
Figure 4.38 Comparison of “weighted thickness” and “weighted properties” cross‐sectional area.
117
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 10
0.5
1
1.5
hhole/H
I x,av
g/I x,g (
stro
ng a
xis)
weighted thicknessweighted properties
Figure 4.39 Comparison of “weighted thickness” and “weighted properties” strong axis moment of inertia.
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 10
0.5
1
1.5
hhole/H
I y,av
g/I y,g (
wea
k ax
is)
weighted thicknessweighted properties
Figure 4.40 Comparison of “weighted thickness” and “weighted properties” weak axis moment of inertia.
.7.3.2.3 215BAverage section property calculations for the column - J and Cw
The average J and Cw of the 1200S162‐68 column with circular holes is determined
directly using ABAQUS and then compared to the “weighted properties” and
118
“weighted thickness” predictions in this section. The differential equation for torsion is
defined as (Timoshenko 1961):
3
3
dxdEC
dxdGJT w
ββ−= , (4.12)
where β is the angle of twist of the cross section and G is the shear modulus of steel
(G=11346 ksi in this case). Eq. 1233H(4.12) is used in conjunction with static ABAQUS analyses
(not eigenbuckling!) to solve for Javg and Cw,avg of the column as hhole/H varies from 0.10 to
0.90. Javg is calculated by applying a unit twist at the end of the column about the gross
cross‐section shear center while keeping the opposite column end fixed against twist as
shown in 1234HFigure 4.41. If both ends of the column are free to warp, the variation in twist
along the column is constant as shown in 1235HFigure 4.42 and warping resistance does not
contribute to the resulting torsion (d3β/dx3=0). The variation in twist was not sensitive to
increasing hole diameter in this case, and therefore the line shown in 1236HFigure 4.42 is the
same regardless of hole diameter. The twist β along the column is measured in
ABAQUS as the relative rotation of the flange‐web corners. The twist magnitude along
the column length remains unchanged with increasing hhole/H. Javg for the warping free
column is calculated by rearranging Eq. 1237H(4.12):
o
oavg
LGTJ
β= (4.13)
βo, G, and L are known and To is the torque resulting from the unit rotation βo, which is
read directly from ABAQUS.
119
βoΤο
shear center reference node (determined with gross cross section)
1 (x)
2
3
Cross-section doffixed in 2 and 3 (warping free)
Cross-section kinematically restrained to shear center reference node in 2, 3 (warping free)
A
A SECTION A-A
Node at centerline of web fixed in 1 to prevent rigid body motion
Figure 4.41 ABAQUS boundary conditions for warping free and applied unit twist at x=0 in. and warping free but rotation restrained at x=100 in.
0 10 20 30 40 50 60 70 80 90 1000
0.2
0.4
0.6
0.8
1
1.2
1.4
distance along column, in.
β/β 0, n
orm
aliz
ed a
ngle
of t
wis
t
Figure 4.42 Angle of twist decreases linearly in the SSMA 1200S162‐68 column with warping free end conditions.
The resulting Javg from ABAQUS is compared against the “weighted properties” and
“weighted thickness” calculations of Javg. (Note that the “weighted properties” Javg is
calculated with Jnet from 1238HFigure 4.37 using the CUFSM section property calculator and
120
assuming the thickness is zero at the hole). It is clear from 1239HFigure 4.43 that the
“weighted properties” calculation of Javg is most consistent with Javg derived from
ABAQUS.
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 10
0.5
1
1.5
hhole/H
J avg/J
g
weighted thicknessweighted propertiesABAQUS
Figure 4.43 The “weighted properties” approximation for Javg matches closely with the ABAQUS prediction for the SSMA 12S00162‐68 column with holes
The ABAQUS boundary conditions are now modified such that warping is
restrained at the fixed column end as shown in 1240HFigure 4.44. A unit twist, βo, is again
applied at the free end, and the resulting angle of twist β along the length of the column
is measured in ABAQUS. Because of the warping‐fixed end condition, β is nonlinear
along the length of the column and warping resistance contributes to the torsional
stiffness of the column. Since the distribution of β along the column is not influenced by
hhole/H as observed in ABAQUS, an indirect solution of Cw,avg as a function of Cw,g can be
derived:
121
( )
( )0
0
,,
,
=−
=−=
xdxdGJT
xdxdGJT
CC
ggo
avgo
gw
avgw
β
β
, (4.14)
where for each ABAQUS model (hhole/H=0.10 to 0.90), the torque To associated with the
unit twist βo is read directly from ABAQUS and dβ/dx(x=0) is calculated from 1241HFigure 4.45.
As was the case for the warping free case in 1242HFigure 4.42, the variation in twist was not
sensitive to increasing hole diameter and therefore the line shown in 1243HFigure 4.45 is the
same regardless of hole diameter. The influence of holes on the variation in twist is
expected to be more pronounced as column length decreases relative to hole length.
Future research is planned to evaluate the influence of member length on the torsional
properties of columns with holes.
βoΤο
shear center reference node (determined with gross cross section)
1 (x)
2
3
Cross-section doffixed in 1,2, and 3 (warping fixed)
Cross-section kinematically restrained to shear center reference node in 2, 3 (warping free)
A
A SECTION A-A
Figure 4.44 ABAQUS boundary conditions for warping free and applied unit twist at x=0 in. and warping fixed and rotation restrained at x=100 in.
122
0 10 20 30 40 50 60 70 80 90 1000
0.2
0.4
0.6
0.8
1
1.2
1.4
distance along column, in.
β/β 0, n
orm
aliz
ed a
ngle
of t
wis
t
Figure 4.45 Angle of twist is nonlinear along the SSMA 1200S162‐68 column with warping fixed end conditions at x=100 in.
1244HFigure 4.46 demonstrates that the “weighted properties” and “weighted thickness”
approximations both overestimate Cw,avg when compared to the ABAQUS derived Cw,avg
demonstrating that neither is an accurate predictor of Cw,avg. A modified version of the
“weighted properties” approximation is also plotted, where Cw,net is taken equal to zero
instead of Cw,net taken from the results in 1245HFigure 4.37. This assumption for Cw,net is
motivated by the idea that the hole separates the cross section into two pieces, where
each piece on its own contributes minimally to warping resistance. This modified
“weighted properties” approximation results in a conservative lower bound on Cw,avg
which is useful from a design perspective until more accurate approximations are
developed.
123
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 10
0.5
1
1.5
hhole/H
Cw
,avg
/Cw
,g
weighted thicknessweighted propertiesderived from ABAQUSweighted properties Cw ,net=0
Figure 4.46 Comparison of “weighted thickness” and “weighted properties” approximations to the ABAQUS derived warping torsion constant Cw,avg.
.7.3.2.4 216BComparison of prediction accuracy between methods 1246HFigure 4.47 compares the weak‐axis flexural critical elastic buckling load of the
1200S162‐68 column calculated with the “weighted thickness” and “weighted
properties” prediction methods to ABAQUS eigenbuckling results. The ABAQUS
calculation of Pcre is systematically 10% lower than the prediction method (even for a
column without holes), which results from the assumption of a rigid cross‐section in the
classical stability equations. The column cross‐section as modeled in ABAQUS is
allowed to change shape along the length, resulting in a lower axial stiffness of the
column. (The reduction in Pcre was also confirmed in CUFSM, which like ABAQUS,
accounts for plate‐type deformations in elastic buckling calculations.) Beyond this
systematic difference, both approximate methods are accurate predictors of Pcre for
124
hhole/h≤0.60 and the “weighted properties” method remains accurate for even larger holes.
The prediction of weak‐axis flexure Pcre using the net section properties from 1247HFigure 4.37
is also plotted in 1248HFigure 4.47, demonstrating a conservative alternative to the “weighted
properties” and “weighted thickness” methods.
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 10
0.5
1
1.5
hhole/H
Pcr
e,AB
AQU
S/Pcr
e,pr
edic
tion
weighted thicknessweighted propertiesnet section
Difference is caused by classical rigid cross-section assumption which increases axial column stiffness when compared to ABAQUS
Figure 4.47 Comparison of “weighted thickness” and “weighted properties” prediction methods for the SSMA 1200S162‐68 weak‐axis flexural buckling mode. Predictions using net section properties are also
plotted as a conservative benchmark.
1249HFigure 4.48 compares the “weighted thickness” and “weighted properties” methods
to ABAQUS results for the second global mode, flexural‐torsional column buckling. The
accuracy of the prediction methods decrease with increase hhole/H for both methods,
confirming what was observed in 1250HFigure 4.46, that the weighted approximations for Cw
are not accurate representations of the average warping torsion stiffness, especially as
hhole/H becomes large. Warping torsion dominates over St. Venant torsion in this mode
since both weighted average methods predict similar variations in Pcre, even though J
varies between the two methods (see 1251HFigure 4.43). The “weighted properties” method
125
with Cw,avg replaced with Cw,avg predicted in ABAQUS (see 1252HFigure 4.46) accurately predicts
Pcre until hhole/H exceeds 0.80, although this method may not be practical from a design
perspective since it requires thin shell FE analysis. The modified “weight properties”
approach, calculated assuming Cw,net=0, is shown to be more accurate than using just the
net section properties and is a conservative method for predicting Pcre of flexural‐
torsional buckling modes. Future research is planned to develop a mechanics‐based
approximation for the average Cw of a column including the influence of holes.
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 10
0.5
1
1.5
hhole/H
Pcr
e,AB
AQU
S/Pcr
e,pr
edic
tion
weighted thicknessweighted propertiesnet sectionweighted properties, ABAQUS Cw ,avg
weighted properties, Cw ,net=0
Figure 4.48 Comparison of “weighted thickness” and “weighted properties” prediction methods for the SSMA 1200S162‐68 flexural‐torsional column buckling mode. Predictions using net section properties are
also plotted as a conservative benchmark.
126
4.3 33BElastic buckling of beams with holes
3.1 74BAnalysis of existing tests on beams
A column experiments database was assembled in Section 1253H4.2.6.1 to serve as a
resource in the development and validation of the Direct Strength Method for columns
with holes. In this section, a similar database is developed that summarizes the elastic
buckling properties and tested strengths of cold‐formed steel beam experiments with
holes. This database is used in 1254HChapter 8 when developing and verifying DSM for
beams with holes.
The beam experiments considered in this study were conducted by Shan,
LaBoube, Schuster, and Batson in the early 1990’s and consist of three separate test
sequences (Batson 1992; Schuster 1992; Shan and LaBoube 1994). Test Sequences 1 and 2
were performed at the University of Missouri‐Rolla (UMR) and Test Sequence 3 was
executed at the University of Waterloo. Each specimen is made up of two C‐sections
oriented toe‐to‐toe as depicted in 1255HFigure 4.49. 3/4”x3/4”x1/8” aluminum angles connect
the top and bottom flanges of the two channels with one self‐drilling screw per flange.
The angles provide a closed beam section that prevents lateral‐torsional buckling of the
individual channels.
3.1.1 130BMember and hole dimension notation
Beam cross‐section and hole dimension notation is presented in 1256HFigure 4.50. The
C‐section inside corner radii are assumed to equal twice the measured thickness of the
specimen. Two hole shapes were considered in the experiments, an industry standard
127
slotted hole and a tri‐slotted hole with the curved hole ends replaced by triangular tips.
The holes are centered in the web and are mechanically punched at 24 inches on center
longitudinally with a hole at the center of the span.
3/4”x3/4”x1/8” angle(top and bottom)
self-drilling screw (typ.)
Cee channel (typ.)
Figure 4.49 Cross section of beam specimen showing aluminum strap angles connected to C‐ flanges
B21
H2
D21
r
hhole
t
D22
B22
B11
H1
D1
h
D2
B12
Channel 1 Channel 2
6 in.
Slotted hole
rhole
Lhole
Tri-slotted hole
hhole/2
Figure 4.50 C‐section and hole dimension notation
3.1.2 131BTested boundary conditions and loading
All beam specimens are tested as simply supported in four‐point bending to
create a region of constant moment between the load points at the center of the beam.
128
The point loads are applied through stub channels attached to the beam webs with self‐
drilling screws. The stub channels prevent web crippling by distributing the
concentrated load and by restraining the web. Lateral bracing is provided in the vicinity
of the constant moment region by struts connecting the top flange aluminum angles to a
reaction frame. The ends of the beam specimens are laterally braced by vertical rollers.
A summary of the beam test setup is provided in 1257HFigure 4.51.
129
X
L
X
P P
12” spacing (typ.) Test Sequence 1, 36” spacing (typ.) Test Sequence 2
3/4”x3/4”x1/8” angle connected with one screw to each flange (top and bottom)
6’-0”
a
a
24” hole spacing (typ.)
Section a-a
3/4”x3/4”x1/8” angle
Channel 1 Channel 2
** * * * * * * *
lateral brace point at angle (typ.)
Figure 4.51 Experiment test setup with hole spacing, location of lateral bracing, spacing of aluminum angle straps, and load points
130
3.1.3 132BFinite Element Modeling
The elastic buckling properties of the 72 beam specimens are obtained with
eigenbuckling analyses in ABAQUS (ABAQUS 2004). All beams are modeled with S9R5
reduced integration nine‐node thin shell elements. Refer to Section 1258H2.1 for a detailed
discussion of the S9R5 element. Cold‐formed steel material properties are assumed as
E=29500 ksi and ν=0.3.
Special care is taken to simulate the experimental boundary conditions when
modeling in ABAQUS. The simple supports with vertical roller restraints, the
aluminum angle straps connecting the top and bottom channel flanges, the lateral
bracing of the top flange in the constant moment region, and the application of load
through the webs and are all considered. 1259HFigure 4.52 summarizes the ABAQUS
boundary condition assumptions.
Beam end restrained in 2 and 3 (v, w=0)
Beam end restrained in 2 and 3 (v, w=0)
Bottom flange restrained in 1 at support (u=0)
* **
**
**
**
Restrain node at midline of top flange in 3 (w=0) (Typ.)
Rigid body connection between top (and bottom) flange midline nodes (Typ.)
Figure 4.52 Finite element model boundary conditions for beam eigenbuckling analyses
To simulate the simply supported conditions with vertical rollers, the ends of the
beams are modeled as warping free except for the bottom flange at one end which is
131
restrained to prevent longitudinal rigid body movement. The channel cross‐sections are
restrained from vertical and lateral translation at both beam ends.
A rigid body restraint is used to model the connectivity between the top and
bottom C‐section flanges provided by the aluminum angle straps connected with self‐
drilling screws. 1260HFigure 4.53 demonstrates how each angle is modeled as a rigid body
made up of one midline flange node from each channel. The rigid body definition
requires that the motion (both translational and rotational) of the two nodes is governed
by a single reference node, in this case the midline flange node of Channel 1. The
formulation allows for rigid body motion but requires that the relative position of the
two nodes remains constant. A disadvantage of this rigid body restraint is that flange
movements are only restrained at the midline node and do not simulate contact between
the channel flange and aluminum angle, which sometimes results in distortional
buckling modes that would not be physically possible.
1 inch mesh spacing (typ.)
Rigid body connection between top (and bottom) flange midline nodes
Rigid body reference node
Channel 1
Channel 2 Figure 4.53 Channel and hole meshing details and modeling of aluminum angle straps
Element meshing is performed with a custom‐built Matlab program written by the
author (See 1261HAppendix A). 1262HFigure 4.53 provides an example of a typical FE mesh, where
132
the holes are defined with a series of element lines radiating from the opening. 1263HFigure
4.54 provides a close‐up view of the rounded corner meshing of the channels. Two
elements model the rounded corners here because S9R5 elements have quadratic shape
functions which allow the initial curved geometry. Refer to Section 1264H2.2 for more
information on modeling rounded corners with S9R5 elements.
2-ABAQUS S9R5 finite elements used for rounded corners of channels (max element aspect ratio of 16 to 1)
Figure 4.54 ABAQUS meshing details for C‐section rounded corners
Concentrated loads are applied to the beam specimens through vertical stub
channels connected to the beam webs with self‐drilling screws. To simulate the
distribution of the load into a channel web, the concentrated load is applied as a group
of web point loads in ABAQUS. 1265HFigure 4.55 demonstrates how the concentrated loads
are applied to the beam webs in the finite element models. The web local buckling
restraint (essentially doubling up of the web at the loading point) provided by the stub
channels is not modeled in ABAQUS because it was observed to have a negligible
influence on the elastic buckling behavior in the relatively long constant moment regions
of the beams.
133
The transfer of load from the stub channels through the self-drilling screws is simulated as a series of web point loads in ABAQUS
Stub channel
Applied load
Figure 4.55 Modeling of the beam concentrated loads in ABAQUS
3.1.4 133B Elastic buckling results and mode definitions
The beam specimen elastic buckling modes were reviewed in ABAQUS by the
author to identify the pure local (L) and distortional (D) buckling modes as well as any
mixed elastic modes created by the addition of web holes. Lateral‐torsional buckling is
restrained by the top flange lateral bracing and aluminum angle straps (see 1266HFigure 4.51),
although other possible global (G) buckling modes are possible as discussed in Section
1267H4.3.1.4.3.
The mode identification process for beams with holes is guided by the
experiences obtained in Section 1268H4.2.4 for cold‐formed steel compression members with
holes. C‐section columns with web holes exhibited unique mixed buckling modes
where distortion of the flanges near the hole mixes with local buckling (LH mode). In
134
this beam study mixed local‐distortional modes are again observed, as well as local web
hole modes initiated by the compression component of the stress gradient from bending.
4.3.1.4.1 195BLocal Buckling
Slotted holes in the beam specimen webs initiate unique local buckling modes
and reduce the critical elastic local buckling moment Mcrl in most cases. The shallow
beam specimen without holes (nominal depth of 2.5 inches) in 1269HFigure 4.56 exhibits local
buckling in both the top flange and web. The addition of slotted web holes creates a
new local buckling mode, the LH2 mode. The LH2 mode occurs when the strip of web
above the hole buckles in two half‐waves. This mode occurs because the fundamental
local buckling half‐wavelength of the cross‐section, Lcrl, is less than the length of the
hole. The critical elastic buckling moment for the LH2 mode is 8 percent less than that of
the pure L mode, suggesting that this local hole mode may influence the load‐
deformation response of the beam.
1270HFigure 4.57 compares the elastic buckling behavior of a slightly deeper beam
(nominal depth of 3.625 inches) with and without holes. The addition of slotted web
holes again creates the LH2 mode with a critical elastic buckling moment that is 17
percent less than the pure L mode. 1271HFigure 4.58, 1272HFigure 4.59, and 1273HFigure 4.60 summarize
the influence of slotted holes on the local buckling behavior of deeper beams with
nominal heights of 6 inches, 8 inches, and 12 inches respectively. The LH mode is
identified in these deeper beam depths as the buckling of the strip of web above the hole
into a single half‐wave. The LH2 mode is observed in the 6 in. and 8 in. deep beams but
135
with a higher critical elastic buckling moment that the LH mode. The LH2 mode is not
observed for the 12 in. deep specimen since Lcrl for this specimen exceeds the length of
the hole.
LH2Mcrl/Myg=0.77
LMcrl/Myg=0.83
LMcrl/Myg=0.82
Figure 4.56 Local buckling modes for specimen 2B,20,1&2(H) with and without holes
LMcrl/Myg=3.00
LH2Mcrl/Myg=2.49
LMcrl/Myg=3.02
Figure 4.57 Local buckling modes for specimen 3B,14,1&2(H) with and without holes
136
LMcrl/Myg=1.05
LH2Mcrl/Myg=0.87
LMcrl/Myg=1.07
LHMcrl/Myg=0.75
Figure 4.58 Local buckling modes for specimen 6B,18,1&2(H) with and without holes
LMcrl/Myg=0.78
LH2Mcrl/Myg=0.72
LMcrl/Myg=0.79
LHMcrl/Myg=0.63
Figure 4.59 Local buckling modes for specimen BP‐40(H) with and without holes
137
LMcrl/Myg=0.96
LMcrl/Myg=0.89
LHMcrl/Myg=0.85
Figure 4.60 Local buckling modes for specimen 12B,16,1&2(H) with and without holes
4.3.1.4.2 196BDistortional Buckling
1274HFigure 4.61 compares the influence of slotted web holes on the distortional
buckling of a shallow beam specimen (nominal height of 2.5 inches). For the specimen
with holes, a unique DH+L mode is observed with a critical elastic buckling moment 20
percent less that the pure D mode. This mode has similar characteristics to the LH mode
in beams (see Section 1275H4.3.1.4.1), especially the buckling of the strip above the hole into
one half‐wave. The DH+L mode is expressed more as a distortional mode though
because the compression flange is wide relative to the unstiffened strip. The D mode
without holes becomes a mixed distortional‐local mode (D+L) when holes are added,
although the critical elastic buckling moment is not significantly affected in this case.
This specimen is sensitive to mixing of local and distortional modes because of the
138
relatively thin steel sheet thickness t of 0.0346 inches. It is also noted that the hole has
only a small influence on the pure D mode half‐wavelength for this specimen.
The DH distortional buckling mode at the hole is also observed for a slightly
deeper beam specimen (nominal height of 3.625 inches) in 1276HFigure 4.62. The sheet
thickness for these channels is roughly double that of the previously discussed specimen
(t=0.71 inches) and the hole depth is unchanged. Mcrl is higher than Mcrd because of the
increased thickness, resulting in DH and D modes without local buckling interaction
when the slotted holes are present. The critical elastic buckling moment of the DH mode
is 13 percent less than that of the pure D mode.
DMcrd/Myg=1.08Half wavelength=13 inches
DH+LMcrd/Myg=0.83
D+LMcrd/Myg=1.03Half wavelength=14 inches
Figure 4.61 Distortional buckling modes for specimen 2B,20,1&2(H) with and without holes
139
DMcrd/Myg=2.31Half wavelength=12 inches
DHMcrd/Myg=2.00
DMcrd/Myg=2.31Half wavelength=12 inches
Figure 4.62 Distortional buckling modes for specimen 3B,14,1&2(H) with and without holes
1277HFigure 4.63 and 1278HFigure 4.64 compare the influence of slotted web holes on beams
with nominal heights of six inches and eight inches respectively, both having a steel
sheet thickness of t=0.047 inches. For these specimens the unstiffened strip buckling
mode above the hole is identified as LH buckling (see 1279HFigure 4.58, 1280HFigure 4.59) instead of
DH buckling because the majority of the buckling deformation occurs in the web. The
similarities between the LH and DH modes can make them difficult to classify in some
cases. Research on a mechanics‐based modal identification method is underway.
DMcrd/Myg=1.56Half wavelength=12 inches
DMcrd/Myg=1.56Half wavelength=12 inches
Figure 4.63 Distortional buckling modes for specimen 6B,18,1&2(H) with and without holes
140
DMcrd/Myg=1.02Half wavelength=12 inches
DMcrd/Myg=1.00Half wavelength=12 inches
Figure 4.64 Distortional buckling modes for specimen BP5‐40(H) with and without holes
D+LMcrd/Myg=1.03Half wavelength=11 inches
DMcrd/Myg=1.02Half wavelength=12 inches
Figure 4.65 Distortional buckling modes for specimen 12B,16,1&2(H) with and without holes
The distortional buckling modes of the deepest beam specimen considered in this
study (nominal depth of 12 inches) are provided in 1281HFigure 4.65. Identifying the
distortional buckling modes for the channels making up this beam are inherently
challenging because even for a member without holes, there is not a clear distinction
between the L and D modes. The critical elastic buckling moments for a single C‐section
from the beam cross section are provided at various half‐wavelengths from a finite strip
analysis (CUFSM) in 1282HFigure 4.66 (including the modal participation factors calculated
with the constrained finite strip method). Only one minimum exists, suggesting that the
modes at or near the minimum buckling load are a mixture of L and D modes. The most
suitable mode identified by the author (for the specimen without a hole) in 1283HFigure 4.65
alternates between larger distortional half‐waves and shorter local buckling half‐waves
in the constant moment region of the channels. For the specimen with the web holes in
141
1284HFigure 4.65, the local half‐waves are not present and the mode resembles more of a
“pure” D mode. The DH mode is not observed for this specimen, which is consistent
with the buckling behavior of stiffened elements in bending (see 1285HFigure 3.25).
Unstiffened strip buckling (the plate mode that is hypothesized to initiate the DH mode
in beams) does not occur when hhole/h is small.
L
This plot summarizes the modal participation (L, D, G, O) as a function of half-wavelength
D
Figure 4.66 Elastic buckling curve for 12” deep specimen with modal participation summarized, note that
selected L and D are mixed local‐distortional modes
4.3.1.4.3 197BGlobal buckling
Lateral‐torsional buckling is a common global (G) elastic buckling mode in
beams, although this mode is eliminated for the specimens considered here by
connecting the two C‐sections toe‐to‐toe with aluminum angles and by providing lateral
bracing at the compression flange in the constant moment region of the beams (see
1286HFigure 4.51). Twisting of an individual channel about its longitudinal axis is still
possible though, even with the top flange restrained. 1287HFigure 4.67 depicts the potential
twisting mode. CUFSM is used to conservatively quantify the elastic buckling moment
for this mode, and it is determined that Mcre is more than ten times the yield moment My
142
for the specimens in this study. Since the Mcre does not influence the DSM prediction as
long as Mcre > 2.78 My , the global twisting mode is not summarized in the database.
Lateral bracing of top flange
Figure 4.67 Possible global buckling mode occurs about the compression flange lateral brace point
3.1.5 134BElastic buckling database for beams with holes
1288HTable 4.11 summarizes the dimensions and material properties of each channel
making up the beam (Channel 1 and Channel 2), including cross section and hole
dimensions, tested ultimate point load Ptest (for each channel) and ultimate moment Mtest
(for each channel), tested specimen yield stress Fy, specimen yield moment My,g
(calculated with the gross cross‐section), and My,net (calculated with the net cross‐
section). Fy varies from 22.0 ksi to 93.3 ksi with a mean of 48.6 ksi and standard
deviation of 14 ksi. This large variation in yield stress was somewhat unexpected.
ABAQUS eigenbuckling results are summarized in 1289HTable 4.12 for each channel
considering the experiment boundary conditions both with and without holes. These
results are used in Section 1290H4.3.1.6 to evaluate the influence of holes on Mcrl and Mcrd. The
CUFSM elastic buckling results are also provided, including the fundamental
distortional half‐wavelength Lcrd, which are used in Section 1291H4.3.1.7 to evaluate the
influence of experiment boundary conditions on Mcrl and Mcrd and the distortional half‐
wavelength.
143
1292HTable 4.10 presents the cross‐section parameter ranges of the beam C‐sections
contained in the experiment database. All of the beam specimens have cross‐section
dimensions that meet the DSM prequalification standards for ultimate strength
prediction summarized in 1293HTable 4.9 (AISI‐S100 2007). Four of the beam specimens
exceed the yield stress prequalification limit of Fy<70 ksi.
Table 4.9 DSM prequalification limits for beam C‐sections
Beam parameter DSM prequalification limit
Web slenderness H/t<321Flange slenderness B/t<75 Lip slenderness 0<D/t<34Web / flange 1.5<H/B<17Lip / flange 0<D/B<0.70Yield stress Fy<70 ksi
Table 4.10 Parameter ranges for beam specimens with holes D/t H/t B/t H/B D/B hhole/h F y (ksi)
min 5.5 40.5 16.3 1.5 0.18 0.13 22.0max 22.1 257.1 58.3 7.7 0.42 0.67 93.3
144
Table 4.11 Beam experiment cross‐section dimensions, material properties, and tested strengths
1 2 1 2Test
Sequence L shear span t E nu Hole Type L hole h hole r hole H 1 H 2 B 11 B 21 B 12 B 22 D 11 D 21 D 12 D 22 r F y M y,g M y,g M y,net M y,net P test M test
in. in. in. ksi in. in. in. in. in. in. in. in. in. in. in. in. in. in. ksi k*in k*in k*in k*in kips k*inShan and LaBoube 1994 1 2,16,1&2(H) 150.0 39.0 0.062 29500 0.3 Slotted 2.000 0.750 0.375 2.510 2.510 1.610 1.610 1.630 1.610 0.400 0.450 0.420 0.430 0.124 37.23 12.25 12.18 12.18 12.11 1.04 10.1
1 2,20,1&2(H) 150.0 39.0 0.039 29500 0.3 Slotted 2.000 0.750 0.375 2.500 2.480 1.600 1.600 1.600 1.600 0.420 0.410 0.420 0.410 0.078 33.70 7.13 7.04 7.09 7.00 0.46 4.51 2,20,3,4(H) 150.0 39.0 0.039 29500 0.3 Slotted 2.000 0.750 0.375 2.510 2.520 1.590 1.620 1.580 1.600 0.360 0.420 0.470 0.410 0.078 33.70 7.17 7.19 7.14 7.16 0.46 4.51 3,14,1&2(H) 150.0 39.0 0.077 29500 0.3 Slotted 4.000 1.500 0.750 3.680 3.680 1.650 1.640 1.630 1.630 0.570 0.550 0.560 0.520 0.154 63.72 43.11 42.72 42.35 41.96 3.7 36.11 3,14,3&4(H) 150.0 39.0 0.077 29500 0.3 Slotted 4.000 1.500 0.750 3.690 3.690 1.630 1.620 1.640 1.630 0.530 0.530 0.620 0.550 0.154 63.72 43.83 43.07 43.05 42.30 3.54 34.51 3,18,1&2(H) 150.0 39.0 0.044 29500 0.3 Slotted 4.000 1.500 0.750 3.750 3.650 1.560 1.560 1.570 1.580 0.580 0.560 0.580 0.540 0.088 46.92 19.05 18.30 18.74 17.98 1.35 13.21 3,18,3&4(H) 150.0 39.0 0.044 29500 0.3 Slotted 4.000 1.500 0.750 3.650 3.640 1.560 1.580 1.560 1.570 0.560 0.570 0.540 0.540 0.088 46.92 18.18 18.20 17.86 17.88 1.37 13.41 3,20,1&2(H) 150.0 39.0 0.044 29500 0.3 Slotted 4.000 1.500 0.750 3.650 3.710 1.560 1.640 1.550 1.590 0.520 0.560 0.550 0.560 0.088 46.82 18.11 18.90 17.79 18.59 1.35 13.21 3,20,3&4(H) 150.0 39.0 0.044 29500 0.3 Slotted 4.000 1.500 0.750 3.670 3.690 1.560 1.590 1.550 1.610 0.600 0.560 0.520 0.590 0.088 46.82 18.14 18.93 17.83 18.61 1.43 13.91 12,14,1&2(H) 192.0 60.0 0.098 29500 0.3 Slotted 4.000 1.500 0.750 12.080 12.070 1.640 1.630 1.690 1.630 0.690 0.600 0.600 0.620 0.196 35.93 167.26 164.89 167.09 164.73 7.16 107.41 12,14,3&4(H) 192.0 60.0 0.098 29500 0.3 Slotted 4.000 1.500 0.750 12.050 12.000 1.640 1.600 1.670 1.710 0.650 0.640 0.650 0.640 0.196 35.93 167.02 166.45 166.85 166.28 7.50 112.51 12,16,1&2(H) 192.0 60.0 0.055 29500 0.3 Slotted 4.000 1.500 0.750 11.960 11.970 1.570 1.570 1.570 1.560 0.500 0.610 0.520 0.430 0.110 49.11 124.61 123.30 124.48 123.17 4.38 65.71 12,16,3&4(H) 192.0 60.0 0.055 29500 0.3 Slotted 4.000 1.500 0.750 12.070 11.960 1.560 1.570 1.570 1.580 0.420 0.530 0.580 0.530 0.110 49.11 126.99 125.25 126.86 125.12 4.79 71.92 2B,16,1&2(H) 150.0 39.0 0.059 29500 0.3 Slotted 4.000 1.500 0.750 2.460 2.460 1.620 1.630 1.620 1.610 0.470 0.460 0.510 0.510 0.118 53.59 16.68 16.62 15.93 15.88 1.345 13.12 2B,16,3&4(H) 150.0 39.0 0.059 29500 0.3 Slotted 4.000 1.500 0.750 2.470 2.460 1.630 1.620 1.620 1.630 0.470 0.520 0.520 0.460 0.118 53.59 16.80 16.57 16.06 15.84 1.36 13.32 2B,20,1&2(H) 150.0 39.0 0.033 29500 0.3 Slotted 4.000 1.500 0.750 2.420 2.420 1.630 1.640 1.630 1.620 0.420 0.420 0.500 0.500 0.066 67.15 11.99 11.95 11.46 11.43 0.6 5.92 2B,20,3&4(H) 150.0 39.0 0.033 29500 0.3 Slotted 4.000 1.500 0.750 2.420 2.430 1.630 1.640 1.630 1.620 0.420 0.410 0.500 0.500 0.066 67.15 11.99 12.01 11.46 11.49 0.635 6.22 3B,14,1&2(H) 150.0 39.0 0.071 29500 0.3 Slotted 4.000 1.500 0.750 3.650 3.620 1.620 1.660 1.630 1.630 0.540 0.550 0.490 0.500 0.142 81.36 49.80 49.51 48.89 48.61 4.31 42.02 3B,14,3&4(H) 150.0 39.0 0.071 29500 0.3 Slotted 4.000 1.500 0.750 3.640 3.630 1.630 1.620 1.620 1.630 0.540 0.470 0.490 0.540 0.142 81.36 49.49 49.91 48.58 48.99 4.255 41.52 3B,18,1&2(H) 150.0 39.0 0.044 29500 0.3 Slotted 4.000 1.500 0.750 3.610 3.630 1.610 1.650 1.650 1.620 0.510 0.520 0.500 0.500 0.088 53.13 20.81 20.81 20.44 20.45 1.6 15.62 3B,18,3&4(H) 150.0 39.0 0.044 29500 0.3 Slotted 4.000 1.500 0.750 3.620 3.630 1.620 1.660 1.650 1.640 0.500 0.500 0.520 0.520 0.088 53.13 20.98 21.05 20.61 20.68 1.51 14.72 3B,20,1&2(H) 150.0 39.0 0.036 29500 0.3 Slotted 4.000 1.500 0.750 3.610 3.600 1.630 1.620 1.630 1.620 0.460 0.470 0.460 0.470 0.072 63.71 20.36 20.25 20.00 19.89 1.2 11.72 3B,20,3&4(H) 150.0 39.0 0.036 29500 0.3 Slotted 4.000 1.500 0.750 3.610 3.610 1.640 1.630 1.640 1.630 0.460 0.470 0.470 0.470 0.072 63.71 20.48 20.40 20.12 20.04 1.1 10.72 3B,20,5&6(H) 150.0 39.0 0.036 29500 0.3 Slotted 4.000 1.500 0.750 3.600 3.600 1.630 1.630 1.620 1.630 0.460 0.460 0.460 0.470 0.072 63.71 20.22 20.33 19.86 19.97 1.335 13.02 3B,20,1&2(T) 150.0 39.0 0.029 29500 0.3 Tri-slotted 4.500 1.500 --- 3.560 3.570 1.620 1.650 1.680 1.600 0.590 0.640 0.620 0.610 0.058 25.51 6.82 6.68 6.70 6.56 0.425 4.12 3B,20,3&4(T) 150.0 39.0 0.029 29500 0.3 Tri-slotted 4.500 1.500 --- 3.560 3.560 1.620 1.680 1.690 1.610 0.580 0.630 0.620 0.570 0.058 25.51 6.84 6.64 6.73 6.52 0.455 4.42 6B,18,1&2(H) 192.0 60.0 0.046 29500 0.3 Slotted 4.000 1.500 0.750 6.060 6.050 1.620 1.620 1.550 1.550 0.470 0.470 0.500 0.500 0.092 47.17 37.34 37.26 37.14 37.06 1.64 24.62 6B,18,3&4(H) 192.0 60.0 0.046 29500 0.3 Slotted 4.000 1.500 0.750 6.050 6.020 1.620 1.620 1.550 1.550 0.470 0.480 0.500 0.510 0.092 47.17 37.26 37.10 37.06 36.89 1.7 25.52 6C,18,1&2(H) 192.0 60.0 0.048 29500 0.3 Slotted 4.000 1.500 0.750 5.960 5.960 1.980 1.990 1.980 1.990 0.640 0.590 0.590 0.640 0.096 75.08 70.77 71.51 70.43 71.17 3.425 51.42 6C,18,3&4(H) 192.0 60.0 0.048 29500 0.3 Slotted 4.000 1.500 0.750 5.950 5.980 1.970 1.980 1.990 1.980 0.600 0.650 0.640 0.630 0.096 75.08 71.28 71.59 70.94 71.25 3.445 51.72 6D,18,1&2(H) 192.0 60.0 0.046 29500 0.3 Slotted 4.000 1.500 0.750 6.020 6.020 2.420 2.430 2.430 2.430 0.700 0.620 0.620 0.700 0.092 30.77 32.19 32.55 32.06 32.41 1.67 25.12 6D,18,3&4(H) 192.0 60.0 0.046 29500 0.3 Slotted 4.000 1.500 0.750 6.020 6.020 2.430 2.430 2.430 2.430 0.700 0.700 0.610 0.620 0.092 30.77 32.16 32.20 32.02 32.07 1.7 25.52 6B,20,1&2(H) 192.0 60.0 0.033 29500 0.3 Slotted 4.000 1.500 0.750 5.920 5.920 1.630 1.620 1.520 1.530 0.440 0.470 0.440 0.420 0.066 93.26 50.95 50.90 50.66 50.61 1.15 17.32 8A,14,1&2(H) 192.0 60.0 0.074 29500 0.3 Slotted 4.000 1.500 0.750 8.060 8.060 1.380 1.380 1.380 1.380 0.490 0.480 0.410 0.430 0.148 31.04 53.04 53.24 52.88 53.08 3.675 55.12 8A,14,3&4(H) 192.0 60.0 0.074 29500 0.3 Slotted 4.000 1.500 0.750 8.070 8.070 1.380 1.380 1.380 1.380 0.500 0.410 0.410 0.500 0.148 31.04 53.16 53.93 53.00 53.77 3.7 55.52 8A,14,5&6(H) 192.0 60.0 0.074 29500 0.3 Slotted 4.000 1.500 0.750 8.070 8.070 1.370 1.380 1.380 1.370 0.410 0.500 0.490 0.410 0.148 31.04 53.77 53.02 53.61 52.86 3.64 54.62 8A,14,7&8(H) 192.0 60.0 0.065 29500 0.3 Slotted 4.000 1.500 0.750 8.030 8.030 1.390 1.390 1.390 1.400 0.430 0.480 0.480 0.450 0.130 56.29 86.17 86.08 85.91 85.83 4.37 65.62 8A,14,9&10(H) 192.0 60.0 0.065 29500 0.3 Slotted 4.000 1.500 0.750 8.040 8.040 1.390 1.380 1.380 1.380 0.460 0.440 0.450 0.480 0.130 56.29 85.70 86.09 85.45 85.83 4.31 64.72 8B,14,1&2(T) 192.0 60.0 0.067 29500 0.3 Tri-slotted 4.500 1.500 --- 8.050 8.050 1.640 1.630 1.640 1.640 0.630 0.640 0.670 0.660 0.134 32.58 58.37 58.25 58.22 58.10 3.225 48.42 8B,14,3&4(T) 192.0 60.0 0.067 29500 0.3 Tri-slotted 4.500 1.500 --- 8.050 8.040 1.640 1.640 1.640 1.640 0.640 0.640 0.660 0.650 0.134 32.58 58.29 58.09 58.14 57.93 3.89 58.42 8B,14,5&6(T) 192.0 60.0 0.065 29500 0.3 Tri-slotted 4.500 1.500 --- 8.020 8.020 1.630 1.640 1.640 1.630 0.640 0.630 0.640 0.630 0.130 53.14 91.54 91.20 91.30 90.95 3.735 56.02 8B,14,7&8(T) 192.0 60.0 0.065 29500 0.3 Tri-slotted 4.500 1.500 --- 8.030 8.030 1.630 1.630 1.630 1.630 0.660 0.610 0.610 0.660 0.130 53.14 91.07 91.71 90.83 91.47 5.375 80.62 8D,14,1&2(T) 192.0 60.0 0.065 29500 0.3 Tri-slotted 4.500 1.500 --- 7.950 7.960 2.480 2.500 2.470 2.490 0.640 0.480 0.470 0.620 0.130 54.71 113.41 116.36 113.16 116.10 5.895 88.42 8D,14,3&4(T) 192.0 60.0 0.065 29500 0.3 Tri-slotted 4.500 1.500 --- 7.950 7.950 2.470 2.490 2.470 2.480 0.660 0.480 0.450 0.610 0.130 54.71 113.04 115.71 112.79 115.45 5.925 88.92 8B,18,1&2(H) 192.0 60.0 0.045 29500 0.3 Slotted 4.000 1.500 0.750 7.950 7.940 1.590 1.580 1.580 1.580 0.470 0.470 0.480 0.470 0.090 72.32 82.33 81.96 82.10 81.73 2.76 41.42 8D,18,1&2(H) 192.0 60.0 0.046 29500 0.3 Slotted 4.000 1.500 0.750 8.000 8.000 2.420 2.450 2.440 2.430 0.610 0.690 0.690 0.620 0.092 22.00 33.87 33.58 33.80 33.51 2.1 31.52 8D,18,3&4(H) 192.0 60.0 0.046 29500 0.3 Slotted 4.000 1.500 0.750 8.000 8.000 2.420 2.450 2.450 2.430 0.600 0.700 0.700 0.600 0.092 22.00 33.97 33.48 33.90 33.41 1.84 27.62 8A,20,1&2(H) 192.0 60.0 0.031 29500 0.3 Slotted 4.000 1.500 0.750 7.930 7.930 1.380 1.390 1.380 1.380 0.410 0.440 0.450 0.430 0.062 37.96 27.95 27.90 27.86 27.81 1.005 15.12 8A,20,3&4(H) 192.0 60.0 0.031 29500 0.3 Slotted 4.000 1.500 0.750 7.930 7.920 1.370 1.380 1.390 1.370 0.450 0.430 0.430 0.440 0.062 37.96 27.94 27.80 27.85 27.71 0.985 14.82 8B,20,1&2(T) 192.0 60.0 0.031 29500 0.3 Tri-slotted 4.500 1.500 --- 7.970 7.970 1.630 1.640 1.630 1.620 0.610 0.610 0.600 0.620 0.062 44.89 37.40 37.47 37.30 37.37 1.37 20.62 8B,20,3&4(T) 192.0 60.0 0.031 29500 0.3 Tri-slotted 4.500 1.500 --- 7.960 7.960 1.630 1.630 1.620 1.630 0.620 0.580 0.580 0.620 0.062 44.89 37.13 37.42 37.03 37.32 1.4 20.62 8B,20,5&6(T) 192.0 60.0 0.031 29500 0.3 Tri-slotted 4.500 1.500 --- 7.950 7.950 1.630 1.630 1.630 1.630 0.610 0.600 0.600 0.610 0.062 44.89 37.27 37.32 37.17 37.22 1.4 21.42 8B,20,7&8(T) 192.0 60.0 0.031 29500 0.3 Tri-slotted 4.500 1.500 --- 7.950 7.950 1.630 1.630 1.640 1.630 0.610 0.620 0.610 0.620 0.062 44.89 37.42 37.41 37.32 37.31 1.4 20.72 8D,20,1&2(T) 192.0 60.0 0.043 29500 0.3 Tri-slotted 4.500 1.500 --- 7.940 7.940 2.490 2.450 2.450 2.490 0.640 0.590 0.590 0.640 0.086 38.59 54.64 55.30 54.52 55.18 2.5 38.12 8D,20,3&4(T) 192.0 60.0 0.043 29500 0.3 Tri-slotted 4.500 1.500 --- 7.940 7.940 2.460 2.460 2.440 2.480 0.640 0.590 0.590 0.650 0.086 38.59 54.46 55.30 54.34 55.18 2.7 39.82 8D,20,5&6(T) 192.0 60.0 0.043 29500 0.3 Tri-slotted 4.500 1.500 --- 7.950 7.950 2.490 2.460 2.450 2.480 0.620 0.620 0.620 0.630 0.086 38.59 54.95 55.27 54.83 55.15 2.6 39.02 12B,16,1&2(H) 192.0 60.0 0.060 29500 0.3 Slotted 4.000 1.500 0.750 11.950 11.950 1.630 1.630 1.630 1.630 0.530 0.540 0.520 0.530 0.120 60.64 169.78 170.15 169.61 169.98 6.5 97.32 12B,16,3&4(H) 192.0 60.0 0.060 29500 0.3 Slotted 4.000 1.500 0.750 11.980 12.020 1.630 1.630 1.620 1.630 0.470 0.500 0.550 0.530 0.120 60.64 170.43 171.33 170.26 171.16 6.4 96.62 12B,16,5&6(H) 192.0 60.0 0.060 29500 0.3 Slotted 4.000 1.500 0.750 11.960 11.970 1.630 1.630 1.630 1.630 0.510 0.500 0.510 0.520 0.120 60.64 169.55 169.96 169.38 169.79 6.4 95.82 12B,16,7&8(H) 192.0 60.0 0.060 29500 0.3 Slotted 4.000 1.500 0.750 11.970 11.960 1.630 1.630 1.620 1.630 0.480 0.550 0.560 0.490 0.120 60.64 170.58 169.32 170.41 169.15 6.7 100.0
Schuster 1992 3 BP4—40(H) 168.0 48.0 0.047 29500 0.3 Slotted 4.02 1.5 0.750 7.99 7.99 1.61 1.61 1.61 1.61 0.47 0.47 0.47 0.47 0.094 38.87 46.76 46.76 46.64 46.64 3.2 37.93 BP5—40(H) 168.0 48.0 0.047 29500 0.3 Slotted 4.02 1.5 0.750 7.99 7.99 1.61 1.61 1.61 1.61 0.47 0.47 0.47 0.47 0.094 38.87 46.76 46.76 46.64 46.64 3.1 36.83 BP6—40(H) 168.0 48.0 0.047 29500 0.3 Slotted 4.02 1.5 0.750 7.99 7.99 1.61 1.61 1.61 1.61 0.47 0.47 0.47 0.47 0.094 38.87 46.76 46.76 46.64 46.64 3.2 38.23 BP7—65(H) 168.0 48.0 0.047 29500 0.3 Slotted 4.53 2.48 0.750 7.99 7.99 1.58 1.58 1.58 1.58 0.47 0.47 0.47 0.47 0.094 38.87 46.33 46.33 45.74 45.74 3.1 37.73 BP8—65(H) 168.0 48.0 0.047 29500 0.3 Slotted 4.53 2.48 0.750 7.99 7.99 1.61 1.58 1.61 1.58 0.47 0.47 0.47 0.47 0.094 38.87 46.76 46.33 46.18 45.74 3.2 38.23 BP9—65(H) 168.0 48.0 0.047 29500 0.3 Slotted 4.53 2.48 0.750 7.99 7.99 1.61 1.58 1.58 1.58 0.47 0.47 0.47 0.47 0.094 49.02 58.55 58.43 57.82 57.69 3.2 38.23 CP4—40(T) 168.0 48.0 0.048 29500 0.3 Tri-slotted 4.65 1.69 --- 7.99 7.99 1.58 1.58 1.58 1.58 0.51 0.51 0.51 0.51 0.096 49.02 60.19 60.19 59.95 59.95 3.5 41.43 CP5—40(T) 168.0 48.0 0.048 29500 0.3 Tri-slotted 4.65 1.69 --- 7.99 7.99 1.58 1.61 1.58 1.58 0.51 0.51 0.51 0.51 0.096 49.02 60.19 60.32 59.95 60.08 3.3 39.43 CP6—40(T) 168.0 48.0 0.048 29500 0.3 Tri-slotted 4.65 1.69 --- 8.03 8.03 1.61 1.61 1.58 1.58 0.53 0.51 0.51 0.51 0.096 49.02 60.81 60.75 60.57 60.52 3.5 41.63 CP7—65(T) 168.0 48.0 0.048 29500 0.3 Tri-slotted 4.61 2.52 --- 7.99 7.99 1.61 1.61 1.61 1.61 0.51 0.51 0.51 0.51 0.096 49.02 60.75 60.75 59.96 59.96 3.4 41.33 CP8—65(T) 168.0 48.0 0.048 29500 0.3 Tri-slotted 4.61 2.52 --- 8.03 7.99 1.58 1.61 1.58 1.61 0.51 0.51 0.51 0.53 0.096 49.02 60.62 60.98 59.84 60.19 3.4 40.93 CP9—65(T) 168.0 48.0 0.048 29500 0.3 Tri-slotted 4.61 2.52 --- 7.99 7.99 1.61 1.61 1.63 1.61 0.53 0.51 0.51 0.51 0.096 49.02 61.09 60.75 60.30 59.96 3.4 40.8
Experimental Results
Material Properties
Member Length, Load Location and ThicknessStudy and Specimen Name
Yield Stress and Moment of InertiaChannel Channel Cross Section DimensionsHole Dimensions
145
Table 4.12 Beam experiment elastic buckling results
Test Sequence M cr l ,LH M cr l ,LH2 M cr l ,L M crd,DH M crd,D L crd * M cr l ,LH M cr l ,LH2 M cr l ,L M crd,DH M crd,D L crd * M cr l M crd L crd * M cr l M crd L crd * M cr l M crd L crd M cr l M crd L crd
k*in k*in k*in k*in k*in in. k*in k*in k*in k*in k*in in. k*in k*in in. k*in k*in in. k*in k*in in. k*in k*in in.Shan and LaBoube 1994 1 2,16,1&2(H) 65.6 --- 67.0 50.3 38.5 12.0 66.3 --- 67.0 50.3 41.7 12.0 67.56 38.67 12.0 67.97 41.93 12.0 67.34 36.38 8.90 68.45 40.36 10.80
1 2,20,1&2(H) 17.0 --- 17.0 --- 15.1 12.0 16.8 --- 17.0 17.5 14.6 12.0 17.41 15.11 12.0 17.26 14.63 12.0 17.23 14.46 12.90 17.07 14.02 12.801 2,20,3,4(H) 16.9 --- 17.1 --- 13.0 12.0 17.0 --- 17.2 --- 15.0 12.0 17.29 13.02 12.0 17.27 15.02 12.0 17.13 12.57 10.80 17.13 14.38 13.101 3,14,1&2(H) --- 153.7 193.1 115.7 106.5 12.0 --- 153.7 193.1 115.7 103.9 12.0 192.81 108.11 12.0 192.81 105.69 12.0 192.63 105.73 10.80 193.11 102.82 10.801 3,14,3&4(H) --- 149.2 194.3 109.9 101.8 12.0 --- 149.2 194.2 110.0 101.8 12.0 192.98 103.74 12.0 193.09 103.74 12.0 192.17 100.37 10.90 193.72 100.85 10.901 3,18,1&2(H) --- 29.6 38.6 28.3 39.3 12.0 --- 29.6 38.6 28.3 36.9 12.0 38.82 39.60 12.0 38.46 37.25 12.0 38.47 34.97 16.10 37.99 33.35 15.701 3,18,3&4(H) --- 29.7 38.2 28.5 36.6 12.0 --- 29.8 38.2 28.5 37.1 12.0 38.07 37.12 12.0 38.07 37.83 12.0 38.02 33.32 15.70 37.70 33.41 15.601 3,20,1&2(H) --- 29.4 37.7 28.0 33.8 12.0 --- 30.2 37.6 28.1 36.4 12.0 37.81 33.96 12.0 37.48 37.25 12.0 37.81 31.56 15.70 37.13 32.52 15.901 3,20,3&4(H) --- 29.8 38.3 29.2 40.3 12.0 --- 29.8 38.4 28.5 36.9 12.0 38.28 40.56 12.0 38.22 37.44 12.0 38.26 35.31 15.80 37.67 33.19 15.801 12,14,1&2(H) --- --- 248.8 --- 250.9 12.0 --- --- 234.0 --- 234.6 12.0 279.02 259.43 12.0 271.46 239.70 12.0 280.76 236.79 11.50 271.13 236.79 11.501 12,14,3&4(H) --- --- 241.7 --- 243.5 12.0 --- --- 236.6 --- 238.1 12.0 268.64 250.06 12.0 260.27 243.99 12.0 276.49 247.64 11.50 272.01 240.79 11.501 12,16,1&2(H) --- --- 44.6 --- 51.0 12.0 --- --- 46.8 --- 56.4 12.0 47.71 51.22 12.0 49.93 57.30 12.0 48.73 50.74 11.40 49.73 60.03 13.801 12,16,3&4(H) --- --- 41.9 --- 44.0 12.0 --- --- 45.0 --- 46.1 12.0 45.90 45.66 12.0 48.11 50.85 12.0 47.84 46.59 13.90 47.84 52.46 11.402 2B,16,1&2(H) --- 53.8 57.3 43.5 39.9 14.0 --- 53.2 57.3 43.0 39.9 14.0 57.15 47.86 12.0 57.01 46.82 12.0 57.50 37.15 10.60 56.85 36.11 10.602 2B,16,3&4(H) --- 53.4 58.5 43.0 41.6 14.0 --- 54.6 58.5 47.0 41.6 14.0 57.44 48.17 12.0 58.26 53.24 12.0 57.27 37.08 10.60 58.09 40.76 12.702 2B,20,1&2(H) 9.3 9.3 9.9 10.0 12.4 14.0 9.3 9.3 9.9 10.0 12.4 14.0 9.95 12.96 12.0 9.90 12.96 12.0 9.85 9.82 15.10 9.77 9.75 15.102 2B,20,3&4(H) 9.3 9.3 9.7 10.2 12.3 14.0 9.3 9.3 9.7 10.0 11.9 14.0 9.96 13.09 12.0 9.92 12.75 12.0 9.85 9.82 15.10 9.79 9.58 15.202 3B,14,1&2(H) --- 124.1 150.2 99.5 114.9 12.0 --- 124.1 149.3 99.5 113.5 12.0 149.82 114.50 12.0 149.82 114.50 12.0 152.67 85.95 10.80 149.19 85.39 12.902 3B,14,3&4(H) --- --- 151.0 102.0 114.3 12.0 --- --- 150.1 91.7 100.5 12.0 150.47 115.07 12.0 150.13 101.52 12.0 151.76 85.53 10.70 148.57 75.98 10.702 3B,18,1&2(H) 28.3 29.5 36.7 35.7 38.9 12.0 28.4 29.5 36.7 35.7 39.1 12.0 36.63 41.75 12.0 36.63 41.75 12.0 36.82 30.06 15.50 36.32 29.93 15.602 3B,18,3&4(H) 28.1 29.3 36.4 35.3 40.5 14.0 28.1 29.3 36.4 35.3 40.4 14.0 36.36 41.23 15.0 36.36 40.72 15.0 36.62 29.50 15.50 35.99 28.88 15.602 3B,20,1&2(H) 15.6 16.2 20.2 19.7 24.4 14.0 15.6 16.2 20.2 19.7 24.6 14.0 20.17 24.92 15.0 20.17 25.40 15.0 19.95 17.75 15.50 20.05 18.16 15.502 3B,20,3&4(H) 15.6 16.1 20.1 19.6 23.0 16.0 15.6 16.1 20.1 19.6 23.3 16.0 20.08 24.86 15.0 20.08 25.31 15.0 19.84 17.67 15.50 19.99 18.11 15.502 3B,20,5&6(H) 15.6 16.1 20.1 19.6 23.0 16.0 15.6 16.1 20.1 19.6 23.0 16.0 20.09 24.91 15.0 20.09 24.91 15.0 19.92 17.71 15.50 19.91 17.72 15.502 3B,20,1&2(T) 8.6 8.7 10.4 --- 15.9 25.0 8.7 8.7 10.4 --- 15.9 25.0 10.41 20.76 20.0 10.41 20.76 20.0 10.43 14.39 22.30 10.26 15.02 22.302 3B,20,3&4(T) 8.6 8.7 10.5 --- 16.1 25.0 --- 8.7 10.2 --- 15.8 25.0 10.24 20.28 20.0 10.34 20.48 20.0 10.43 14.18 18.40 10.08 14.61 22.302 6B,18,1&2(H) 28.0 32.6 39.9 --- 58.3 12.0 28.0 32.6 39.9 --- 58.3 12.0 39.38 58.49 12.0 39.38 58.49 12.0 39.13 42.30 14.80 39.16 42.28 14.802 6B,18,3&4(H) 28.1 32.6 39.9 --- 58.6 12.0 28.1 32.6 40.1 --- 58.6 12.0 39.39 58.87 12.0 39.64 58.87 12.0 39.16 42.28 14.80 39.37 42.90 14.702 6C,18,1&2(H) 36.4 41.9 50.6 59.8 73.2 20.0 35.7 41.2 49.9 58.8 73.2 20.0 49.97 65.68 19.0 49.25 77.15 19.0 49.74 54.18 21.20 48.85 50.96 17.602 6C,18,3&4(H) 35.5 41.1 49.8 57.3 74.9 23.0 36.3 41.9 50.6 59.7 77.5 23.0 49.16 78.83 19.0 49.91 85.21 19.0 48.80 52.00 21.20 49.63 54.97 21.302 6D,18,1&2(H) 35.0 39.7 47.3 47.3 59.2 24.0 34.0 38.8 44.5 46.6 59.2 24.0 44.76 64.75 24.0 44.05 63.81 24.0 44.52 47.53 25.80 43.55 42.89 25.802 6D,18,3&4(H) 34.9 39.6 45.3 47.4 66.7 27.0 34.9 39.6 45.3 47.4 66.7 27.0 44.85 68.02 24.0 44.85 68.02 24.0 44.52 47.37 25.80 44.49 47.37 25.802 6B,20,1&2(H) 10.7 12.3 15.1 --- 28.1 12.0 10.8 12.4 15.3 --- 28.1 12.0 14.93 28.30 12.0 15.07 28.30 12.0 14.82 19.61 17.40 14.96 20.78 17.402 8A,14,1&2(H) 103.0 117.5 131.3 133.0 129.6 6.0 103.0 116.6 132.6 130.8 128.3 6.0 129.97 128.02 6.0 127.47 126.88 6.0 127.62 122.79 9.30 127.00 120.90 9.302 8A,14,3&4(H) 104.6 117.9 134.2 133.0 130.2 6.0 98.3 109.6 129.0 120.7 120.3 6.0 130.42 128.44 6.0 124.20 116.81 6.0 128.58 124.46 9.30 127.39 109.29 9.302 8A,14,5&6(H) 98.0 109.1 128.6 120.3 119.9 5.0 104.7 118.0 133.1 133.2 130.2 5.0 123.80 116.30 6.0 128.76 128.47 6.0 126.98 108.98 9.30 128.71 124.56 9.302 8A,14,7&8(H) 69.2 81.0 88.5 90.9 86.8 5.0 71.5 78.5 91.1 91.4 90.2 5.0 86.93 85.59 5.0 88.37 96.81 5.0 85.18 83.61 9.20 88.03 90.23 11.202 8A,14,9&10(H) 70.8 80.2 90.3 90.6 89.0 6.0 69.5 78.7 88.7 88.6 87.2 6.0 87.44 95.68 7.0 85.87 91.40 7.0 87.17 87.94 9.30 85.54 84.53 9.302 8B,14,1&2(T) 86.1 98.7 110.6 --- 152.3 12.0 86.1 98.7 110.8 --- 152.3 12.0 107.87 151.94 11.0 107.98 151.94 11.0 107.41 123.59 16.30 107.55 124.50 16.302 8B,14,3&4(T) 86.5 99.2 111.1 --- 153.3 12.0 86.5 99.2 111.3 --- 153.3 12.0 108.35 152.82 11.0 108.50 152.82 11.0 107.91 124.67 16.30 108.08 124.73 16.302 8B,14,5&6(T) 79.2 90.7 101.7 --- 137.1 12.0 79.2 90.7 101.7 --- 137.1 12.0 99.22 140.34 11.0 99.22 140.34 11.0 98.79 116.37 16.20 98.95 115.58 16.202 8B,14,7&8(T) 80.1 91.6 102.6 --- 140.3 14.0 78.3 89.7 100.6 --- 136.8 10.0 100.03 141.76 12.0 98.12 141.76 12.0 99.81 118.65 16.20 97.56 112.93 13.502 8D,14,1&2(T) 96.1 109.5 121.2 --- 137.4 24.0 90.2 103.9 115.2 --- 125.6 13.0 118.25 137.30 24.0 112.42 126.65 14.0 118.12 108.07 19.40 111.09 83.24 16.102 8D,14,3&4(T) 96.7 110.0 121.9 --- 138.3 24.0 90.2 103.9 115.2 --- 126.2 12.0 118.89 143.78 24.0 112.38 127.17 14.0 118.74 111.45 23.40 111.08 83.37 16.102 8B,18,1&2(H) 25.9 29.4 32.8 --- 46.7 10.0 25.8 29.3 32.7 --- 46.7 10.0 31.96 47.69 11.0 31.88 47.69 11.0 31.78 41.89 13.30 31.73 41.80 13.302 8D,18,1&2(H) 34.0 38.3 42.3 --- 76.7 27.0 35.2 39.5 43.5 --- 76.7 27.0 41.15 78.20 24.0 42.37 78.20 24.0 40.75 50.55 23.60 42.16 56.35 28.502 8D,18,3&4(H) 33.9 38.2 42.2 --- 75.9 24.0 35.3 39.6 43.7 --- 75.9 24.0 41.02 77.69 24.0 42.50 77.69 24.0 40.58 49.76 23.60 42.33 56.95 28.502 8A,20,1&2(H) 8.0 9.1 10.1 --- 16.0 12.0 8.1 9.1 10.3 --- 16.8 10.0 9.86 17.06 10.0 10.02 17.06 10.0 9.80 15.47 13.30 9.98 16.63 13.302 8A,20,3&4(H) 8.1 9.2 10.3 --- 16.7 10.0 8.1 9.1 10.2 --- 16.7 10.0 9.99 17.28 11.0 9.95 17.28 11.0 9.94 16.75 13.30 9.91 16.19 13.302 8B,20,1&2(T) 8.9 10.1 11.4 --- 30.4 17.0 8.9 10.2 11.4 --- 30.4 17.0 11.06 34.95 18.0 11.08 34.95 18.0 11.00 23.66 23.50 11.02 23.69 23.502 8B,20,3&4(T) 9.0 10.2 11.4 --- 28.6 14.0 8.8 10.0 11.3 --- 27.0 12.0 11.11 33.01 18.0 10.96 31.93 18.0 11.07 23.97 23.50 10.88 22.79 19.402 8B,20,5&6(T) 8.9 10.1 11.4 --- 27.9 14.0 8.9 10.1 11.3 --- 28.2 14.0 11.07 32.28 16.0 11.03 32.71 16.0 11.01 23.66 23.40 10.96 23.38 23.402 8B,20,7&8(T) 8.9 10.1 11.3 --- 28.3 13.0 8.9 10.1 11.4 --- 28.3 13.0 11.05 33.66 16.0 11.08 33.66 16.0 10.99 23.64 23.40 11.02 23.93 23.402 8D,20,1&2(T) 28.3 32.1 35.6 53.6 62.6 24.0 27.5 31.3 34.8 53.6 62.6 24.0 33.85 64.10 26.0 34.64 64.10 26.0 34.47 45.07 28.20 33.49 42.24 23.402 8D,20,3&4(T) 28.1 31.9 35.4 56.5 62.6 24.0 27.6 31.4 34.9 56.5 62.6 24.0 34.45 64.07 26.0 33.93 64.07 26.0 34.29 45.48 28.20 33.56 42.15 23.402 8D,20,5&6(T) 28.1 31.9 35.3 56.6 63.3 24.0 27.8 31.6 35.1 56.6 63.3 24.0 34.43 65.10 24.0 34.16 65.10 24.0 34.19 44.07 28.30 33.87 44.53 23.402 12B,16,1&2(H) 59.2 --- 62.0 --- 71.6 11.0 59.2 --- 62.2 --- 71.6 11.0 63.23 72.18 11.0 63.49 72.30 11.0 62.52 71.58 16.60 62.75 72.24 16.602 12B,16,3&4(H) 57.6 --- 60.4 --- 69.0 11.0 58.5 --- 61.2 --- 70.6 11.0 61.49 71.40 11.0 62.15 70.50 11.0 60.45 67.40 16.60 61.49 69.30 16.602 12B,16,5&6(H) 59.0 --- 61.6 --- 75.6 11.0 58.4 --- 61.3 --- 70.3 11.0 62.74 68.34 11.0 62.41 68.34 11.0 61.98 70.20 16.60 61.63 69.43 16.602 12B,16,7&8(H) 57.9 --- 60.7 --- 69.6 11.0 59.9 --- 62.5 --- 73.6 11.0 61.83 73.80 12.0 63.79 73.80 12.0 60.88 68.02 16.60 63.16 73.02 16.60
Schuster 1992 3 BP4—40(H) 29.3 33.5 37.3 --- 47.2 12.0 29.3 33.5 37.3 --- 47.2 12.0 36.39 47.48 11.0 36.39 47.48 11.0 36.22 46.03 13.40 36.22 46.03 13.403 BP5—40(H) 29.3 33.5 37.3 --- 47.2 12.0 29.3 33.5 37.3 --- 47.2 12.0 36.39 47.48 11.0 36.39 47.48 11.0 36.22 46.03 13.40 36.22 46.03 13.403 BP6—40(H) 29.3 33.5 37.3 --- 47.2 12.0 29.3 33.5 37.3 --- 47.2 12.0 36.39 47.48 11.0 36.39 47.48 11.0 36.22 46.03 13.40 36.22 46.03 13.403 BP7—65(H) 26.9 30.8 37.1 --- 47.0 12.0 26.9 30.8 37.1 --- 47.0 12.0 36.12 47.00 12.0 36.12 47.00 12.0 35.95 45.72 13.40 35.95 45.72 13.403 BP8—65(H) 27.4 30.8 37.4 --- 47.6 12.0 27.0 30.8 37.0 --- 45.9 12.0 36.44 47.68 12.0 36.09 47.03 12.0 35.95 45.72 13.40 35.95 45.72 13.403 BP9—65(H) 27.4 31.2 37.4 --- 47.7 12.0 27.0 30.8 37.1 --- 47.1 12.0 36.49 47.73 12.0 36.13 47.09 12.0 35.95 45.72 13.40 35.95 45.72 13.403 CP4—40(T) 30.2 34.9 39.9 --- 51.7 11.0 30.2 34.9 39.9 --- 51.7 11.0 38.88 52.75 11.0 38.88 52.75 11.0 38.71 51.31 13.40 38.71 51.31 13.403 CP5—40(T) 30.3 35.0 39.9 --- 52.6 12.0 30.7 35.3 40.3 --- 53.3 12.0 38.90 52.84 11.0 39.28 53.55 11.0 39.10 51.56 16.20 39.10 51.56 16.203 CP6—40(T) 31.0 35.6 40.5 --- 54.8 12.0 30.7 35.3 40.2 --- 53.2 12.0 39.48 55.34 11.0 39.19 53.43 11.0 39.03 51.52 16.20 39.03 51.52 16.203 CP7—65(T) 30.7 35.2 40.3 --- 53.5 12.0 30.7 35.2 40.3 --- 53.5 12.0 39.17 53.34 12.0 39.17 53.34 12.0 38.99 51.50 16.20 38.99 51.50 16.203 CP8—65(T) 30.4 34.8 39.8 --- 52.8 12.0 30.8 35.2 40.3 --- 53.6 12.0 38.72 52.57 12.0 39.18 53.44 12.0 38.92 51.46 16.20 38.92 51.46 16.203 CP9—65(T) 31.1 35.4 40.5 --- 53.6 12.0 30.7 35.2 40.2 --- 53.6 12.0 39.43 55.28 12.0 39.14 53.37 12.0 38.99 51.50 16.20 38.99 51.50 16.20
L crd * approximated in ABAQUS
Channel 1 Channel 2
Study and Specimen NameABAQUS elastic buckling with holes, experiment boundary conditions
Channel 1 Channel 2 Channel 1
CUFSM elastic buckling, no hole
Channel 2
ABAQUS elastic buckling WITHOUT holes, experiment boundary conditions
146
3.1.6 135BInfluence of holes on beam local and distortional critical elastic buckling loads
4.3.1.6.1 198BLocal buckling
The ABAQUS local buckling eigenbuckling results for each beam specimen C‐section
(Channel 1 and Channel 2) with holes is compared to the same beam specimen but
without holes in 1294HFigure 4.68. The variation in Mcr for the LH, LH2, and L modes (see
Section 1295H4.3.1.4.1 for definition) with hole size to flat web depth is highlighted in 1296HFigure
4.68a. The LH mode (buckling of the compressed unstiffened strip above a hole) is
observed only when 0.20<hhole/h<0.40, and is always the lowest buckling mode when it
exists. As hhole/h exceeds 0.40 the lowest mode switches to the LH2 mode. This trend
occurs because as h decreases, the local buckling half‐wavelength decreases causing
multiple half‐waves to form in the unstiffened strip at the hole. When hhole/h<0.20 the
unstiffened strip above the hole is relatively stiff (i.e., deep relative to hole length) and
plate buckling controls as the lowest local buckling mode. The minimum Mcr for the LH,
LH2, and L is plotted in 1297HFigure 4.68b exhibits a similar trend to that observed for
stiffened elements in bending (see 1298HFigure 3.26a), where the maximum hole influence
occurs when hhole/h is between 0.30 and 0.40. Unstiffened strip buckling (LH and LH2) of
full members controls for hhole/h exceeding 0.50 which is also consistent with the behavior
of a stiffened element in bending (and different from a column with holes, where web
local buckling occurs away from the hole for large hhole/h). The presence of the C‐section
flanges reduces the magnitude of the hole influence in a full member when compared to
147
a stiffened element, which is consistent with similar observations for compression
members (see 1299HFigure 4.4).
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 10
0.5
1
1.5
hhole/h
Mcr
l ,ABA
QU
S ho
le/M
crl ,A
BAQ
US
no h
ole
LHLH2L
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 10
0.5
1
1.5
hhole/h
Mcr
l ,ABA
QU
S ho
le/M
crl ,A
BAQ
US
no h
ole
LH controlsLH2 controlsL controls
Figure 4.68 Influence of holes on beam specimen Mcrl (Channel 1 and Channel 2 plotted) considering (a) all
local buckling modes and (b) the lowest local buckling mode
4.3.1.6.2 199BDistortional buckling
The ABAQUS distortional buckling eigenbuckling results for each beam specimen C‐
section (Channel 1 and Channel 2) with holes is compared to the same beam specimen
but without holes in 1300HFigure 4.69. The variation in Mcr for the DH and D modes (see
Section 1301H4.3.1.4.2 for definition) with hole size to flat web depth is highlighted in 1302HFigure
4.69a. The DH mode is often the lowest distortional mode in 1303HFigure 4.69b, especially
when hhole/h is between 0.20 and 0.40. This mode is initiated by unstiffened strip buckling
and is related to the LH mode, and therefore its maximum influence in this region is
expected.
The ratio of web depth to flange width is an important parameter to consider when
differentiating between the LH and DH modes for beams with holes. The DH mode is
most prevalent in the range 2<H/B<6 as shown in 1304HFigure 4.70. As the beam depth
increases relative to flange width (H/B>6) the distortional tendency associated with
148
unstiffened strip buckling decreases and the DH mode transitions to the LH (or LH2)
mode.
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 10
0.5
1
1.5
hhole/h
Mcr
d,AB
AQU
S ho
le/M
crd,
ABAQ
US
no h
ole
DHD
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1
0
0.5
1
1.5
hhole/h
Mcr
d,AB
AQU
S ho
le/M
crd,
ABAQ
US
no h
ole
DH controlsD controls
Figure 4.69 Influence of holes on beam specimen Mcrd (Channel 1 and Channel 2 plotted) as a function of hole depth to flat web depth considering (a) all distortional buckling modes and (b) the lowest distortional
buckling mode
0 1 2 3 4 5 6 7 80
0.5
1
1.5
H/B
Mcr
d,AB
AQU
S ho
le/M
crd,
ABAQ
US
no h
ole
DHD
0 1 2 3 4 5 6 7 8
0
0.5
1
1.5
H/B
Mcr
d,AB
AQU
S ho
le/M
crd,
ABAQ
US
no h
ole
DH controlsD controls
Figure 4.70 Influence of holes on beam specimen Mcrd (Channel 1 and Channel 2 plotted) as a function of web depth to flange width considering (a) all distortional buckling modes and (b) the lowest distortional
buckling mode
3.1.7 136BInfluence of experiment boundary conditions on beam local and distortional critical elastic buckling loads
4.3.1.7.1 200BLocal buckling
The influence of experiment boundary conditions on the elastic buckling
behavior is evaluated by comparing the ABAQUS critical elastic buckling moment
149
Mcrl (without holes) of each C‐section making up the beam specimens to the local
buckling moment determined with the finite strip software CUFSM. Since the finite
strip method considers elastic buckling of each channel individually under a constant
moment, the comparison of ABAQUS and CUFSM results isolate the influence of the
aluminum angle straps at the top and bottom flanges, as well as the lateral bracing and
the application of the constant moment as a series of point loads in the experiments. The
experiment loading and boundary conditions have a minimal influence on Mcrl for the
specimens considered in this study as shown in 1305HFigure 4.71. This result is consistent
with the local buckling mode shapes in Section 1306H4.3.1.4.1, where it was observed that the
formation of local buckling half‐waves in the constant moment region were unimpeded
by the angle straps.
0 50 100 150 200 250 3000
0.5
1
1.5
H/t
Mcr
l ,no
hole
, ABA
QU
S/Mcr
l ,CU
FSM
Figure 4.71 Influence of test boundary conditions on Mcrl
150
4.3.1.7.2 201BDistortional buckling
The influence of the experiment boundary conditions on the distortional
buckling behavior is evaluated by comparing the critical elastic buckling moment Mcrd
(without holes) from the ABAQUS eigenbuckling analyses to the buckling moment of
each channel individually determined with the finite strip software CUFSM. The
comparison of ABAQUS and CUFSM results isolates the influence of the aluminum
angle straps, lateral bracing and the load application method on the critical elastic
moment results. The experiment test conditions provide a significant boost to Mcrd as
shown in 1307HFigure 4.72b, which is hypothesized to be related to the restrained distortional
buckling caused by the aluminum angle straps. This hypothesis is supported by existing
research on unrestrained elastic distortional beam buckling (no compression flange
connections), which observed similar CUFSM and ABAQUS eigenbuckling results (Yu
and Schafer 2006). The pure D distortional half‐wavelengths approximated from
ABAQUS (for specimens without holes) in 1308HFigure 4.72b are often shorter relative to the
predicted half‐wavelengths from a finite strip analysis. This trend is a direct result of
the angle spacing (12” on center for Test Sequences 1 and 3, 6” for Test Sequence 2),
which is less than the fundamental Lcrd for many of the C‐sections. The change in half‐
wavelength away from the natural half‐wavelength of the distortional mode increases
the critical elastic buckling moment. This boost in Mcrd decreases with increasing H/B as
shown in 1309HFigure 4.72a, since the fundamental Lcrd also decreases as beam web depth
increases relative to flange width.
151
0 1 2 3 4 5 6 7 80
0.2
0.4
0.6
0.8
1
1.2
1.4
1.6
H/B
Mcr
d,no
hol
e, A
BAQ
US/M
crd,
CU
FSM
0 1 2 3 4 5 6 7 80
0.5
1
1.5
H/B
L crd,
ABAQ
US
no h
ole/L
crd,
CU
FSM
Figure 4.72 Influence of test boundary conditions on (a) Mcrd and (b) on the distortional half‐wavelength
The boost in Mcrd from the restraint of the beam compression flanges exhibits a
linear trend when plotted against the ratio of Lcrd (from CUFSM) versus the restraint
spacing Sbrace in 1310HFigure 4.73. A linear equation is fit to this trend, resulting in a useful
approximation of the restraint boost:
85.015.0 +⎟⎟⎠
⎞⎜⎜⎝
⎛=
brace
crdboost S
LD , 1≥brace
crd
SL
(4.15)
0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 50
0.2
0.4
0.6
0.8
1
1.2
1.4
1.6
1.8
2
Lcrd,CUFSM/Sbrace
Mcr
d,AB
AQU
S no
hol
e/Mcr
d,C
UFS
M
Beam databaseEq. (4.15)DSM Design Guide
Figure 4.73 Boost in Mcrd from the angle restraints increases as the fundamental distortional half‐wavelength
increases relative to the restraint spacing Sbrace
152
The DSM Design Guide’s suggested modification to Mcrd when L<=Lcrd is also plotted in
1311HFigure 4.73 (AISI 2006, Section 4.2):
( )CUFSMcrdLL
CUFSMcrdcrdCUFSMcrdcrd LLMLLM ,ln,
*, )()( =< , (4.16)
where M*crd is the minimum distortional critical elastic buckling moment read from
CUFSM. L is assumed equal to Sbrace when plotted in 1312HFigure 4.73, i.e. the distortional half‐
wave is assumed to form between the flange braces. The DSM Design Guide prediction
for Mcrd is higher than that proposed by Eq. 1313H(4.16) because for many of the beams in the
ABAQUS‐generated elastic buckling database, Lcrd was shortened but not completely
restrained between braces. On the other hand, the ABAQUS eigenbuckling analyses did
not simulate contact between the angles and the flanges (only the bending and shear
stiffness of the angles), and therefore the actual Mcrd most likely lies between the two
predictions.
3.2 75BApproximate prediction methods for use in design
3.2.1 137BLocal buckling
4.3.2.1.1 202BPrediction method
The approximate method for predicting the local elastic buckling behavior of cold‐
formed steel beams is similar to the method for columns presented in Section 1314H4.2.7.1.1
Local buckling is assumed to occur as the minimum of Mcr of the gross cross‐section (as
calculated in the Direct Strength Method) and local buckling of the compressed
unstiffened strip adjacent to the hole, Mcrh. The method captures the lowest unstiffened
153
strip buckling mode, either the LH or LH2 mode, with the procedure described in 1315HFigure
4.17. When the hole length is longer than the fundamental half‐wavelength of the net
cross‐section Lcrh, then the LH2 mode governs. When the hole length is less than Lcrh, the
LH mode governs.
To predict Mcrh from the net cross‐section in CUFSM, the cross‐section is restrained to
isolate local buckling from distortional buckling as shown in 1316HFigure 4.74. Compressed
corners should be restrain in the direction normal to the neutral axis about which
bending occurs (corners experiencing tension need not be restrained). It is important to
avoid fully restraining a cross‐section element, since this prevents Poisson‐type
deformations and artificially stiffens the cross‐section. The only time both the x and z
directions of a corner should be restrained is if a hole isolates two compressed
intersecting elements (as in the case of a flange hole in a C‐section, see 1317HFigure 4.74a).
Finally, when holes isolate two compressed elements of a cross‐section (similar to the
flange hole in the column hat section, see 1318HFigure 4.16b), the isolated element should be
removed from the cross‐section. This prediction method is validated in the next section
using the beam elastic buckling database developed in Section 1319H4.3.1.5.
154
x Neutral axis
z
Web hole Flange holeab
Tension
Compression
Tension
Compression
Tension
Compression
Neutral axis
Tension
Compression
Restrain isolated compressed corners in x and z
Restrain compressed corners in direction normal to neutral axis (typ.)
Figure 4.74 Guidelines for restraining beam net cross‐sections in the CUFSM local buckling approximate method
4.3.2.1.2 203BMethod verification using elastic buckling database
The finite strip prediction method is used to predict Mcrl for the 144 C‐sections
described in 1320HTable 4.11. These predictions are compared to the ABAQUS eigenbuckling
results from 1321HTable 4.12 (the minimum of L, LH, and LH2 modes), and demonstrates that
the finite strip approximate method is viable and conservative over a wide range of hole
widths and beam depths. A clear transition from L and LH2 buckling to LH buckling
occurs as the C‐sections increases in depth as shown in 1322HFigure 4.75a. This observation is
consistent with finite element eigenbuckling observations (see 1323HFigure 4.56 to 1324HFigure
4.60), where as beam depth increases the half‐wavelength of the net‐section increases
beyond the length of the hole, resulting in a switch from unstiffened strip buckling in
two half‐waves (LH2) to one half‐wave (LH). The mean and standard deviation of the
ABAQUS to predicted ratio for Mcrl are 1.14 and 0.16 respectively.
155
0 50 100 150 200 250 300
0
0.2
0.4
0.6
0.8
1
1.2
1.4
1.6
1.8
2
H/t
Mcr
l ,ABA
QU
S/Mcr
l ,pre
dict
ed
L predictedLH2 predictedLH predicted
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1
0
0.2
0.4
0.6
0.8
1
1.2
1.4
1.6
1.8
2
hhole/h
Mcr
l ,ABA
QU
S/Mcr
l ,pre
dict
ed
L predictedLH2 predictedLH predicted
Figure 4.75 Comparison of ABAQUS to predicted Mcrl for C‐sections with holes in the beam database as a function of (a) web depth and (b) hole width relative to flat web depth
3.2.2 138BDistortional buckling
4.3.2.2.1 204BPrediction method
The “weighted average” and “mechanics‐based” finite strip methods for predicting
Mcrd of columns with holes introduced in Section 1325H4.2.7.2.1 are employed here to predict
the distortional critical elastic buckling load of cold‐formed steel beams with holes.
These approximate methods are evaluated against the ABAQUS Mcrd (the minimum of
the DH and D modes) from the beam experiment database in 1326HTable 4.12.
4.3.2.2.2 205BMethod verification using elastic buckling database
1327HFigure 4.76 plots Mcrd determined with ABAQUS versus the predictions using the
“weighted‐average” and “mechanics‐based” approximate methods. The ABAQUS Mcrd
(with holes) is multiplied by the ratio of Mcrd from CUFSM to Mcrd from ABAQUS without
holes to eliminate the influence of the boundary conditions and to allow for a consistent
comparison between the ABAQUS results (with only the hole influence) and the
prediction method. The ABAQUS to “mechanics‐based” prediction ratio is more
156
accurate (ABAQUS to predicted mean of 1.04 and standard deviation of 0.02) than the
“weighted‐average” prediction (ABAQUS to predicted mean of 1.10 and standard
deviation of 0.06), which is consistent with the verification study for columns with holes
in Section 1328H4.2.7.2.1.
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 10
0.2
0.4
0.6
0.8
1
1.2
1.4
1.6
1.8
2
hhole/h
Mcr
d,AB
AQU
S/Mcr
d,pr
edic
ted*M
crd,
CU
FSM/M
crd,
ABAQ
US
no h
ole
weighted averagemechanics-based
Figure 4.76 Comparison of “mechanics‐based” and “weighted‐average” prediction methods to ABAQUS results for the distortional buckling load Mcrd of C‐sections with holes in the elastic buckling database
3.2.3 139BGlobal buckling
The “weighted thickness” and “weighted properties” approximate methods
presented in Section 1329H4.2.7.3.1 are now implemented to predict Mcre for a beam with
uniformly spaced holes loaded with a constant moment.
157
4.3.2.3.1 206BDescription of Prediction Method
The “weighted thickness” and “weighted properties” approximates for Iy, J, and Cw
are employed with the classical lateral‐torsional stability equation to predict Mcre of a
beam with holes (Chajes 1974):
⎟⎟⎠
⎞⎜⎜⎝
⎛+= 2
2
LECGJEIM wycre
π. (4.17)
4.3.2.3.2 207BExample and Verification
The long SSMA 1200S162‐68 member evaluated in Section 1330H4.2.7.3.2 as a column is
now evaluated as a beam with a uniform moment along the member to compare
prediction methods to ABAQUS results. The ABAQUS boundary conditions and
applied loading are described in 1331HFigure 4.77. The beam ends are modeled as warping‐
free and the cross‐section at the longitudinal midline is warping‐fixed. Warping at the
member ends is visible in 1332HFigure 4.77 for this lateral‐torsional buckling mode.
1 (x)
2
3
Cross-section dof fixed in 1 and 1 at longitudinal midline (warping fixed)
Cross-section dof fixed in 2 and 3 (warping free)
Cross-section dof fixed in 2 and 3 (warping free)
Moment applied as consistent nodal loads on cross section (typ.)
Warping free end detail
Figure 4.77 ABAQUS boundary conditions and applied loading for an SSMA 1200S162‐68 beam with holes (hhole/h=0.50 shown)
158
1333HFigure 4.78 demonstrates that both the “weighted stiffness” and “weighted
properties” models are accurate predictors of Mcre for hhole/h≤0.50 in this particular case.
For hhole/h>0.50, the reduction in prediction accuracy occurs because the weighted average
approximations for Cw are not consistent with the actual physical behavior (J was shown
to be consistent with the “weight properties” method for calculating section properties
in Section ). If a designer does not know Cw,avg, then using the net section properties
(calculated with CUFSM, see 1334HFigure 4.37) or the “weighted properties” prediction with
Cw, net=0 are both viable options for conservatively predicting Mcre.
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 10
0.5
1
1.5
hhole/h
Mcr
e,AB
AQU
S/Mcr
e,pr
edic
tion
weighted thicknessweighted propertiesnet sectionweighted properties, ABAQUS Cw ,avg
weighted properties, Cw ,net=0
Figure 4.78 Comparison of “weighted thickness” and “weighted properties” prediction methods for the
SSMA 1200S162‐68 lateral‐torsional beam buckling mode. Predictions using net section properties are also plotted as a conservative benchmark.
159
Chapter 5 4BExperiments on cold-formed steel columns with holes
The elastic buckling modes discussed in 1335HChapter 4 and their influence on the load‐
deformation response of cold‐formed steel columns can be readily observed and
quantified with experiments. In this study, 24 cold‐formed steel lipped C‐section
columns with and without slotted web holes are tested to failure. The column lengths
and cross‐section dimensions are specifically chosen to explore the connection between
local, distortional, and global elastic buckling modes, ultimate strength, and the
resulting failure mechanisms. The elastic buckling behavior is evaluated for each
specimen with a finite element eigenbuckling analysis, taking care to accurately simulate
the tested boundary conditions and measured specimen dimensions. These elastic
buckling results are used to provide a means of understanding the varied deformation
response under load. The columns are tested with friction‐bearing boundary conditions
where the ends of each specimen are milled flat and parallel, and bear directly against
160
steel platens. Recommendations are made to advise other researchers on the viability of
the friction‐bearing boundary conditions when testing short and intermediate length
columns.
5.1 34BAcknowledgements
The cold‐formed steel column tests described in this chapter were completed with a
team effort from the individuals below:
Eric Harden Latrobe Hall Machine Shop Walter Krug Maryland Hall Machine Shop Michael Franckowiak Maryland Hall Machine Shop Dr. Rachel Sangree Johns Hopkins Postdoctoral Researcher Jack Spangler Senior Mechanical Engineer – Structures Lab Nickolay Logvinosky Structures Lab Technician Mario Fasano Johns Hopkins Senior Rebecca Pierce Johns Hopkins Freshman Dawneshia Sanders Baltimore Polytechnic Institute Senior Alexander Pei High School Intern Clark Western Building Systems in Dundalk, MD graciously donated the structural
studs.
5.2 35B Testing Program
Twenty‐four cold‐formed steel lipped C‐section columns with and without pre‐
punched slotted web holes were tested to failure. The primary experimental parameters
are column cross‐section, column length, and the presence or absence of slotted web
holes. The specimen naming convention, as it relates to the testing parameters, is
defined in 1336HFigure 5.1.
161
362-1-24-NH
Cross- section type Specimen number within common group (1,2,3)
Nominal specimen length, 24 in. or 48 in.
Specimen with holes (H) or without holes (NH)
No Holes Holes362-1-24-NH 362-1-24-H362-2-24-NH 362-2-24-H362-3-24-NH 362-3-24-H362-1-48-NH 362-1-48-H362-2-48-NH 362-2-48-H362-3-48-NH 362-3-48-H600-1-24-NH 600-1-24-H600-2-24-NH 600-2-24-H600-3-24-NH 600-3-24-H600-1-48-NH 600-1-48-H600-2-48-NH 600-2-48-H600-3-48-NH 600-3-48-H
SSMA 362S162-33
SSMA 600S162-33
Short Column
Intermediate Column
Short Column
Intermediate Column
Figure 5.1 Column testing parameters and naming convention
2.1 76BRationale for selecting specimen dimensions
2.1.1 140BCross‐section types
Two industry standard cross‐sections from the Steel Stud Manufacturers Association
(SSMA 2001), 362S162‐33 and 600S162‐33, were evaluated in this study. The 362S162‐33
cross‐section has a nominal web width of 3.62 in., while the 600S162‐33 web is wider at
6.00 in. Both sections have a 1.62 in. flange and nominal sheet thickness of 0.0346 in.
Specific measured dimensions are provided in Section 1337H5.2.4.
The buckling half‐wavelengths that form along the length of the specimens are cross‐
section dependent, and can be calculated with the semi‐analytical finite strip method
(FSM) (Schafer and Ádàny 2006). FSM assumes simply supported boundary conditions,
and therefore the local and distortional half‐wavelengths for the cross‐sections studied
here, as provided in 1338HTable 5.1, are only a guide as to the expected half‐wavelength in the
fixed‐fixed tests. The FSM half‐wavelengths are still a useful reference when deciding on
specimen lengths (see Section 1339H5.2.1.2) and identifying buckling modes (see Section 1340H5.3.2),
especially as specimen length increases and local and distortional buckling half‐
162
wavelengths converge to the fundamental (simply supported) half‐wavelengths
reported in 1341HTable 5.1.
Table 5.1 FSM local and distortional buckling half‐wavelengths for nominal 362S162‐33 and 600S162‐33 cross‐sections
Local (L) Distortional (D)in. in.
362 2.8 15.4600 4.7 12.2
Cross-sectionElastic buckling half-wavelength
2.1.2 141BColumn lengths
More than 80% of the tested specimens with holes available in the literature are stub
columns, as depicted in the specimen length histogram of tested specimens provided in
1342HFigure 5.2. (The histogram is constructed with the specimens from the elastic buckling
database in Section 1343H4.2.6.) Stub columns accommodate local buckling half‐waves, but
due to their short length, distortional buckling is typically restrained from forming at
relevant stress levels. The specimen lengths selected in this study, a 24 in. short column
and a 48 in. intermediate length column, ensure that at least one distortional half‐wave
and multiple local half‐waves can form along the length of the column (see 1344HTable 5.1).
Further, at least for North American practice, the selected lengths are more typical of the
unbraced length of actual cold‐formed steel columns in an “all‐steel” design with
bridging in place to brace the studs.
163
0 10 20 30 40 50 60 700
2
4
6
8
10
12
14
16
18
20
column specimen length (in.)
num
ber o
f spe
cim
ens
0 10 20 30 40 50 60 700
2
4
6
8
10
12
14
16
18
20
column specimen length (in.)
num
ber o
f spe
cim
ens
Specimens with holes in this study
Stub column tests
0 2 4 6 8 10 12 14 16 18 200
2
4
6
8
10
12
14
16
18
20
L/H
num
ber o
f spe
cim
ens
Figure 5.2 Tested lengths of cold‐formed steel columns with holes as a function of (a) column length L and and (b) L versus out‐to‐out column width H
2.1.3 Hole type and location
One slotted web hole is located at the mid‐height of the short column to evaluate its
influence at the mid‐length of one distortional buckling half‐wave. Two slotted web
holes are oriented in the intermediate length columns with an industry standard spacing
of 24 in. (SSMA 2001). The holes also coincide with the locations where distortional
buckling half‐waves are expected to have their maximum displacement under load. A
typical short column and intermediate length column specimen with slotted holes is
provided in 1345HFigure 5.3.
164
24 in.
Short column Intermediate length column
Figure 5.3 Typical column specimens with slotted holes
2.2 77BColumn test setup
The column tests were performed with the 100 kip capacity two‐post MTS
machine shown in 1346HFigure 5.4. The upper crosshead and lower actuator are fitted with 12
in. x 12 in. x 1 in. thick chrome‐moly 4140 steel platens ground flat and parallel. The
column specimens bear directly on the steel platens as they are compressed. Friction
between the column ends and the steel platens are the only lateral forces that restrain the
column cross‐section under load. An MTS load cell (model number 661‐23A‐02)
measured the applied compressive force on each specimen, and an internal MTS length‐
voltage displacement transducer (LVDT) reported actuator displacement.
165
All column specimens were loaded in displacement control at a constant rate of
0.004 inches per minute. This rate was selected to ensure that the 3 ksi axial stress per
minute upper limit in the Specification for stub column testing would not be exceeded
(AISI‐TS‐2‐02 2001). An MTS 407 controller was used to operate the hydraulic actuator
during the compression tests.
Load Cell
Fixed Crosshead
Position transducers (with magnet tips)
Friction-bearing boundary conditions (specimen bears directly on steel platen)
Figure 5.4 Column test setup and instrumentation
Two Novotechnik T Series position transducers fitted with ball‐jointed magnet
tips measured the east‐west displacements of the specimen flange‐lip intersections at
column mid‐height. Each transducer has a stroke of six inches and is powered by one 9‐
volt battery. The battery strengths were checked periodically to ensure that a drop in
battery charge did not influence the transducer readings. The load cell and transducer
readings are transmitted as voltage to a PC fitted with a National Instruments data
acquisition card. The voltages are then converted to forces and displacements with the
conversion factors summarized in 1347HTable 5.2. All displacement conversion factors were
166
determined by the author with a voltmeter and digital calipers. The data is plotted to
the PC screen and recorded in a text file with a custom LabVIEW program (Labview
2005).
Magnetic tip
Figure 5.5 Novotechnik position transducer with ball‐jointed magnetic tip
Table 5.2 Voltage conversion factors for column test instrumentation Measurement Source Conversion
Tensile Force MTS Load Cell 1 Volt = 1000 lbfActuator Displacement MTS Internal LVDT 1 Volt = 0.300 in.West Flange Displacement Novotechnik Position Transducer 1 Volt = 0.678 in.East Flange Displacement Novotechnik Position Transducer 1 Volt = 0.678 in.
2.3 78BColumn specimen preparation
All column specimens were cut from 8 ft. structural studs using the Central
Machinery 4 ½ inch metal cutting band saw shown in 1348HFigure 5.15. For short columns
without holes, the whole series of specimens (for example 362‐1‐24‐NH, 362‐2‐24‐NH,
and 362‐2‐24‐NH) was cut from a single 8 ft. structural stud. For all other specimen
types, each specimen was cut from its own individual stud. Tensile coupons for
materials testing were obtained from the leftover stud lengths.
167
Figure 5.6 Central Machinery metal band saw used to rough cut column specimens
The specimen ends were milled to ensure flat and parallel bearing surfaces for
testing. The flatness tolerance across the specimen end is recommended as ±0.001 inches
for stub columns and was adopted as the goal for this study (Galambos 1998a). The
short columns were side‐milled with a Fadel computer numerically‐controlled (CNC)
vertical milling machine. The intermediate length columns were too long for the CNC
machine, and were instead side‐milled with a Bridgeport manual milling machine.
During initial trials the milling process caused troublesome vibrations of the specimen.
The large clamping forces required to dampen the vibration also tended to modify the
shape of the C‐section during the milling process. Unsatisfactory flatness results were
obtained in these trials, with flatness variations of up to ±0.010 inches.
The milling procedure was improved by encasing the specimen ends in bismuth
diaphragms before milling as demonstrated in 1349HFigure 5.7. The diaphragms preserved
the undeformed shape of the specimens, dampened vibration during the milling
168
process, and reduced the clamping force required to hold the specimens in place.
Bismuth is a chemical element that is relatively soft compared to steel at room
temperature and melts at 158 degrees Fahrenheit.
Figure 5.7 362S162‐33 short column specimen with bismuth end diaphragms
Liquid bismuth was poured into custom wood forms at the specimen ends. Once
the bismuth was set, the specimen (with bismuth end diaphragms) was positioned in the
milling machine ( 1350HFigure 5.8 through 1351HFigure 5.11). Several passes were made until the
steel cross‐section and bismuth diaphragm were flush. Both column ends were milled
without removing the specimen from the milling table to reduce the chances of
unparallel bearing ends. The bismuth diaphragms were removed from the specimen
with a few taps of a wooden mallet and then melted down for use with the next
specimen. The flatness tolerance of ±0.001 inches was achieved for all but four
specimens (see Section 1352H5.2.4.4, the maximum out‐of‐flatness was +0.003 in.).
169
Figure 5.8 600S162‐33 short column specimen oriented in CNC machine
Figure 5.9 An end mill is used to prepare the column specimens
170
Figure 5.10 The intermediate length specimens were end‐milled in a manual milling machine
Figure 5.11 The specimens are clamped at the webs only to avoid distortion of the cross‐section
171
2.4 79BColumn specimen measurements and dimensions
2.4.1 142BSpecimen reference system and dimension notation All column dimensions are measured with reference to the orientation of the
specimen in the testing machine. The assumed reference system and specimen
dimension notation are provided in 1353HFigure 5.15.
2.4.2 143BCross‐section measurements
The out‐to‐out dimensions of the web, flanges, and lip stiffeners were measured
with digital calipers and aluminum reference plates at the midlength of the specimens.
The measurement procedure for a typical cross‐section is summarized in 1354HFigure 5.12
(specimen setup) and 1355HFigure 5.13 (cross section dimensions). The outside corner radii
were measured using a set of radius gauges with 1/32 in. increments. The cross‐section
dimensions, based on the average of three independent measurements, are provided for
each specimen in 1356HTable 5.3.
172
Check levelness of measuring platform with the angle indicator. The slope perpendicular to the length of the specimen should be as close to zero as possible.
Clamp the specimen to the measuring platform.
Find and mark the longitudinal midline of the specimen.
Figure 5.12 Setup procedure for measuring specimen cross section dimensions
173
Clamp a beveled aluminum plate to the flange. Use the veneer caliper to measure the distance between the edge of the lip and the outside face of the beveled plate. The true dimension (D1 or D2) is then found by subtracting the thickness of the beveled plate from the veneer caliper reading.
Clamp beveled alumninum plates to the lip and web, ofsetting them longituinally by about 1/2 inch. Make sure that the beveled faces are oriented so that they are touching the channel.
Use the extension on the veneer caliper to measure the distance between the outside face of the lip plate and the inside face of the web plate. Make sure that the extension is flush with the flange surface. The true dimension (B1 or B2) is found by subtracting the the thickness of the beveled plate from the veneer caliper reading.
Clamp beveled alumninum plates to each flange, ofsetting them longituinally by about 1/2 inch. Make sure that the beveled faces are oriented so that they are touching the channel.
Use the extension on the veneer caliper to measure the distance between the outside face of one flange plate and the inside face of the other flange plate. Make sure that the extension is flush with the web surface. The true dimension H is found by subtracting the the thickness of the beveled plate from the veneer caliper reading.
Figure 5.13 Procedure for measuring specimen cross‐section dimensions
174
Clamp beveled aluminum plate to flange.
Measure the flange angle with the angle indicator (F1 and F2).
Clamp the beveled aluminum plate to the stiffener lip. Measure the flange angle using the angle indicator (S1 and S2).
Figure 5.14 Procedure for measuring flange‐lip and flange‐web angles
The four corner angles of each C‐section are measured with a digital angle
indicator as demonstrated in 1357HFigure 5.14. The angle indicator has a precision of 0.1
degrees. The flange‐lip angles S1 and S2 are measured at the midlength of the specimens;
the web‐flange angles F1 and F2 are measured at multiple points along the specimen as
denoted in 1358HTable 5.4. The C‐section corner angle magnitudes, based on the average of
two independent measurements, are provided for each specimen in 1359HTable 5.4.
175
Table 5.3 Summary of measured cross section dimensions
H B1 B2 D1 D2 RT1 RT2 RB1 RB2
in. in. in. in. in. in. in. in. in.362-1-24-NH 3.654 1.550 1.621 0.411 0.431 0.188 0.188 0.172 0.188362-2-24-NH 3.712 1.586 1.585 0.416 0.422 0.172 0.203 0.266 0.281362-3-24-NH 3.623 1.677 1.679 0.425 0.399 0.188 0.172 0.281 0.281362-1-24-H 3.583 1.650 1.595 0.430 0.437 0.188 0.203 0.281 0.281362-2-24-H 3.645 1.627 1.593 0.440 0.391 0.188 0.188 0.281 0.281362-3-24-H 3.672 1.674 1.698 0.418 0.426 0.188 0.188 0.266 0.266362-1-48-NH 3.624 1.611 1.605 0.413 0.426 0.172 0.172 0.281 0.281362-2-48-NH 3.624 1.609 1.585 0.407 0.421 0.188 0.172 0.297 0.281362-3-48-NH 3.614 1.604 1.599 0.425 0.401 0.188 0.188 0.266 0.266362-1-48-H 3.622 1.602 1.595 0.420 0.412 0.172 0.172 0.281 0.281362-2-48-H 3.623 1.594 1.610 0.425 0.403 0.172 0.172 0.281 0.281362-3-48-H 3.633 1.604 1.610 0.395 0.432 0.172 0.172 0.281 0.250600-1-24-NH 6.037 1.599 1.631 0.488 0.365 0.172 0.156 0.250 0.203600-2-24-NH 6.070 1.582 1.614 0.472 0.380 0.203 0.203 0.266 0.266600-3-24-NH 6.030 1.601 1.591 0.369 0.483 0.156 0.172 0.266 0.219600-1-24-H 6.040 1.594 1.606 0.484 0.359 0.172 0.172 0.250 0.219600-2-24-H 6.011 1.608 1.602 0.369 0.500 0.172 0.172 0.203 0.234600-3-24-H 6.032 1.606 1.577 0.360 0.478 0.172 0.172 0.250 0.203600-1-48-NH 6.018 1.621 1.609 0.486 0.374 0.172 0.172 0.234 0.219600-2-48-NH 6.017 1.596 1.601 0.482 0.357 0.172 0.172 0.234 0.234600-3-48-NH 6.026 1.585 1.627 0.489 0.338 0.172 0.172 0.266 0.219600-1-48-H 6.010 1.598 1.625 0.480 0.388 0.188 0.156 0.250 0.219600-2-48-H 6.017 1.589 1.607 0.476 0.356 0.172 0.172 0.234 0.234600-3-48-H 6.062 1.632 1.588 0.366 0.480 0.172 0.172 0.219 0.250
Specimen
Section a-a
Front View
South
North
L
a aWest East
D1
B1
H
B2
D2
F1 F2
S1 S2
RB1
RT1 RT2
RB2
b b
W1 W2Section b-b
X
Hole detail
Lhole
hhole rhole=hhole/2
Figure 5.15 Specimen measurement nomenclature
176
Table 5.4 Summary of measured lip‐flange and flange‐web cross section angles
X S1 S2 X F1 F2 X F1 F2 X F1 F2 X F1 F2 X F1 F2in. degrees degrees in. degrees degrees in. degrees degrees in. degrees degrees in. degrees degrees in. degrees degrees
362-1-24-NH 12 12.767 8.367 6 82.600 84.500 12 86.033 86.833 18 84.533 87.000362-2-24-NH 12 11.367 11.567 6 86.800 84.800 12 87.600 85.467 18 86.400 83.700362-3-24-NH 12 9.567 9.433 6 85.700 85.000 12 86.300 85.400 18 85.600 83.000362-1-24-H 12 11.130 10.930 6 83.200 83.970 12 87.600 85.600 18 84.330 86.430362-2-24-H 12 4.367 10.267 6 86.000 85.133 12 86.333 85.167 18 84.400 84.500362-3-24-H 12 10.533 10.833 6 85.200 86.333 12 87.700 86.133 18 87.667 89.033362-1-48-NH 12 7.800 10.100 12 85.100 85.600 18 84.300 85.000 24 85.000 85.600 30 84.000 85.200 36 85.300 85.700362-2-48-NH 12 8.000 10.800 12 85.500 84.900 18 84.800 85.100 24 84.200 84.600 30 84.800 85.300 36 85.200 84.900362-3-48-NH 12 9.100 12.200 12 86.900 84.000 18 85.800 83.900 24 85.300 84.100 30 86.400 83.400 36 86.100 83.700362-1-48-H 12 8.500 9.800 12 86.500 84.800 18 86.600 85.000 24 85.600 84.200 30 85.500 85.100 36 86.400 84.400362-2-48-H 12 8.300 11.200 12 86.800 84.800 18 86.500 84.200 24 85.600 83.800 30 85.500 84.100 36 86.700 83.800362-3-48-H 12 9.700 7.300 12 85.300 85.200 18 84.700 86.100 24 84.100 85.300 30 84.400 84.700 36 85.200 85.000600-1-24-NH 24 1.567 2.133 6 90.567 92.033 12 92.467 93.733 18 91.433 93.767600-2-24-NH 24 1.733 2.333 6 91.000 92.033 12 91.167 94.067 18 91.467 93.333600-3-24-NH 24 -2.167 3.500 6 93.700 89.767 12 94.067 91.033 18 92.733 89.667600-1-24-H 24 0.967 2.033 6 89.000 91.000 12 90.400 92.267 18 91.200 92.600600-2-24-H 24 1.800 1.100 6 94.433 90.900 12 93.233 88.733 18 91.967 89.000600-3-24-H 24 0.100 4.100 6 93.500 90.000 12 93.300 89.300 18 90.100 86.300600-1-48-NH 24 0.167 1.400 12 91.033 92.933 18 90.833 92.700 24 90.600 92.800 30 91.333 92.900 36 91.667 93.200600-2-48-NH 24 2.000 2.367 12 90.767 91.900 18 90.233 92.300 24 89.900 91.867 30 90.967 92.000 36 91.467 92.767600-3-48-NH 24 2.600 2.300 12 90.000 92.100 18 89.200 91.900 24 90.000 92.100 30 90.700 92.600 36 90.900 92.500600-1-48-H 24 2.533 2.100 12 90.933 92.167 18 91.000 92.767 24 90.000 92.633 30 91.000 92.000 36 91.100 92.967600-2-48-H 24 2.400 1.000 12 89.000 90.700 18 89.200 91.000 24 88.900 91.200 30 89.600 91.600 36 90.200 92.200600-3-48-H 24 0.667 3.633 12 93.067 89.400 18 93.000 89.500 24 92.300 89.433 30 93.467 89.900 36 93.467 89.600NOTE: X is the longitudinal distance from the south end of the specimen
Specimen
177
2.4.3 144BSpecimen thickness
All structural studs were delivered by the manufacturer with a zinc outer coating
applied for galvanic corrosion protection. The total zinc thickness (i.e., summation of
the zinc coating thicknesses applied to each side of the steel sheet) and the base metal
thickness (sheet thickness with total zinc coating removed) defined in 1360HFigure 5.16 were
measured for each specimen. The total zinc thickness was used to calculate the
centerline cross‐section dimensions from the out‐to‐out measurements (see Section
1361H5.2.4.2), which were then input along with the base metal thickness into the nonlinear
finite element models discussed in 1362HChapter 7. The base metal thickness was also used to
calculate the steel yield stress provided in Section 1363H5.2.5.
base metal tbare
t1
t2
tzinc=t1+t2
zinc (typ.)
Figure 5.16 Base metal and zinc thickness definitions
Total zinc thickness and base metal thickness were measured for each specimen
from tensile coupons cut from the west flange, east flange, and web of an untested
section of structural stud. The thickness measurements were made to a precision of
0.0001 inches with a digital micrometer fitted with a thimble friction clutch. The
thickness was determined by averaging five measurements taken within the gauge
178
length of the tensile coupon (see 1364HFigure 5.27 for the definition of gauge length). The
base sheet metal thicknesses tbare,w (web), tbare,f1 (west flange), tbare,f2 (east flange) and
corresponding total zinc coating thicknesses tzinc, tzinc,f1, and tzinc,f2 are summarized for each
specimen in 1365HTable 5.5.
Table 5.5 Specimen bare steel and zinc coating thicknesses
tbare,w tzinc,w tbare,f1 tzinc,f1 tbare,f2 tzinc,f2in. in. in. in. in. in.
362-1-24-NH362-2-24-NH362-3-24-NH362-1-24-H 0.0390 0.0030 0.0391 0.0034 0.0391 0.0028362-2-24-H 0.0368 0.0057 0.0390 0.0023 0.0391 0.0034362-3-24-H 0.0394 0.0027 0.0394 0.0018 0.0394 0.0026362-1-48-NH 0.0392 0.0025 0.0393 0.0020 0.0392 0.0020362-2-48-NH 0.0393 0.0025 0.0394 0.0022 0.0393 0.0026362-3-48-NH 0.0389 0.0013 0.0391 0.0009 0.0390 0.0017362-1-48-H 0.0391 0.0019 0.0393 0.0017 0.0394 0.0017362-2-48-H 0.0390 N/M 0.0391 N/M 0.0391 N/M362-3-48-H 0.0401 0.0000 0.0400 0.0000 0.0397 0.0010600-1-24-NH600-2-24-NH600-3-24-NH600-1-24-H 0.0414 0.0042 0.0422 0.0044 0.0428 0.0030600-2-24-H 0.0427 0.0039 0.0384 0.0084 0.0424 0.0042600-3-24-H 0.0429 0.0031 0.0431 0.0026 0.0430 0.0036600-1-48-NH 0.0434 0.0026 0.0436 0.0024 0.0434 0.0028600-2-48-NH 0.0435 0.0017 0.0430 0.0024 0.0430 0.0023600-3-48-NH 0.0436 0.0015 0.0432 0.0021 0.0433 0.0020600-1-48-H 0.0429 0.0022 0.0426 0.0023 0.0429 0.0021600-2-48-H 0.0429 N/M 0.0428 N/M 0.0431 N/M600-3-48-H 0.0430 N/M 0.0434 N/M 0.0430 N/MNOTE: N/M Not measured
0.0438 N/M
0.0368 0.0415N/M
0.0438 N/M 0.0432 N/M
N/M 0.0372
Web West Flange East FlangeSpecimen
N/M
The zinc coating was removed by immersing the tensile coupons in a ferric
chloride bath for 100 minutes. The immersion time was determined with a study of
coupon thickness variation over time for the 362‐2‐24‐H web and the 600‐2‐24‐H west
flange tensile coupons. The coupons were removed from the ferric chloride bath every
10 minutes, cleaned, and then measured. 1366HFigure 5.17 demonstrates that the coupon
thickness converges to a constant value, the base metal thickness, at approximately 100
minutes.
179
The average zinc coating thickness (i.e., average of tzinc, tzinc,f1, and tzinc,f2) for all
specimens was 0.0026 inches using the ferric chloride method described above.
Specimen coating thickness measurements were also made with a Positest DFT digital
thickness gauge (www.defelsko.com) which produced an average coating thickness for
all specimens of 0.0016 in. At the microscopic level, the bonding of the zinc to the steel
substrate results in a gradient from pure zinc to a mixture of steel and zinc (Porter 1991).
This gradient complicates the identification of the non‐structural thickness of the
galvanic coating. The base thickness and coating thickness determined with the ferric
chloride method (as reported in 1367HTable 5.5) are used throughout this thesis. Accurate
identification of the non‐structural and structural contributions of the galvanic coating is
warranted as a topic of future research, especially since the load‐deformation response
and ultimate strength are sensitive to base metal thickness.
0 0.2 0.4 0.6 0.8 10
0.2
0.4
0.6
0.8
1
0 20 40 60 80 100 1200
0.2
0.4
0.6
0.8
1
1.2
1.4
time (minutes)
coup
on th
ickn
ess/
initi
al th
ickn
ess
362-2-24-H Web Coupon600-2-24-H West Flange Coupon
After 100 minutes, zinc coating (tzinc=t1+t2) has been removed with ferric chloride solution
Figure 5.17 Removal of tensile coupon zinc coating as a function of time
180
2.4.4 145BSpecimen end flatness and length After each specimen was saw cut and milled flat, the vertical height gauge (with
a precision of 0.001 inches) shown in 1368HFigure 5.18 was used to measure the specimen
length and flatness. For each specimen, two independent length measurements were
taken at each rounded corner location described in 1369HFigure 5.19. The height gauge and
specimen are placed on the same steel table to ensure that all measurements are made in
the same reference plane. The steel table was checked for flatness with a dial gauge and
precision stand before measurements proceeded. Lengths LRT1, LRT2, LRB1, and LRB2 as
well as the average length L are provided for each specimen in 1370HTable 5.6. The specimen
flatness, defined as the difference between LRT1, LRT2, LRB1, and LRB2 and the average
length L, is reported in 1371HTable 5.7. All but four specimens met the flatness tolerance of
±0.001 inches, with intermediate length column 362‐2‐48‐H having the maximum
deviation of +0.003 inches at LRT2.
North
Figure 5.18 A height gauge is used to measure specimen length
181
LRB1
LRT1 LRT2
LRB2
West East
Figure 5.19 Lengths are measured at the four corners of the C‐section column
Table 5.6 Measured column specimen length Specimen LRT1 LRT2 LRB1 LRB2 L (avg.)
in. in. in. in. in.362-1-24-NH 24.100 24.100 24.098 24.099 24.099362-2-24-NH 24.097 24.098 24.099 24.099 24.098362-3-24-NH 24.097 24.098 24.098 24.099 24.098362-1-24-H 24.100 24.099 24.098 24.100 24.099362-2-24-H 24.097 24.099 24.099 24.100 24.099362-3-24-H 24.099 24.099 24.099 24.100 24.099362-1-48-NH 48.214 48.214 48.214 48.214 48.214362-2-48-NH 48.303 48.300 48.301 48.298 48.301362-3-48-NH 48.192 48.19 48.191 48.189 48.191362-1-48-H 48.217 48.216 48.216 48.216 48.216362-2-48-H 48.232 48.232 48.231 48.231 48.232362-3-48-H 48.196 48.200 48.195 48.198 48.197600-1-24-NH 24.100 24.101 24.099 24.099 24.100600-2-24-NH 24.102 24.104 24.102 24.103 24.103600-3-24-NH 24.100 24.098 24.099 24.099 24.099600-1-24-H 24.102 24.100 24.100 24.101 24.101600-2-24-H 24.098 24.099 24.100 24.100 24.099600-3-24-H 24.101 24.101 24.101 24.100 24.101600-1-48-NH 48.255 48.255 48.255 48.255 48.255600-2-48-NH 48.250 48.250 48.250 48.251 48.250600-3-48-NH 48.295 48.294 48.295 48.294 48.295600-1-48-H 48.089 48.088 48.089 48.088 48.089600-2-48-H 48.253 48.251 48.253 48.253 48.253600-3-48-H 48.061 48.061 48.059 48.059 48.060
182
Table 5.7 Specimen end flatness
Specimen LRT1 LRT2 LRB1 LRB2in. in. in. in.
362-1-24-NH 0.001 0.001 -0.001 0.000362-2-24-NH -0.001 0.000 0.001 0.001362-3-24-NH -0.001 0.000 0.000 0.001362-1-24-H 0.001 0.000 -0.001 0.001362-2-24-H -0.002 0.000 0.000 0.001362-3-24-H 0.000 0.000 0.000 0.001362-1-48-NH 0.000 0.000 0.000 0.000362-2-48-NH 0.002 -0.001 0.001 -0.002362-3-48-NH 0.002 -0.001 0.001 -0.002362-1-48-H 0.001 0.000 0.000 0.000362-2-48-H 0.001 0.001 0.000 0.000362-3-48-H -0.001 0.003 -0.002 0.001600-1-24-NH 0.000 0.001 -0.001 -0.001600-2-24-NH -0.001 0.001 -0.001 0.000600-3-24-NH 0.001 -0.001 0.000 0.000600-1-24-H 0.001 -0.001 -0.001 0.000600-2-24-H -0.001 0.000 0.001 0.001600-3-24-H 0.000 0.000 0.000 -0.001600-1-48-NH 0.000 0.000 0.000 0.000600-2-48-NH 0.000 0.000 0.000 0.001600-3-48-NH 0.001 -0.001 0.001 -0.001600-1-48-H 0.001 0.000 0.001 0.000600-2-48-H 0.001 -0.001 0.001 0.001600-3-48-H 0.001 0.001 -0.001 -0.001
Flatness (Deviation from Average Length)
2.4.5 146BLocation and dimensions of slotted holes The length and width of the slotted holes, Lhole and hhole, were measured to a
precision of 0.001 inches with digital calipers. The east‐west locations of the holes, W1
and W2, were measured by clamping aluminum plates to the outside surface of the
flanges and then using the caliper extension to measure the distance from the edge of the
hole to the aluminum plate. (This process is similar to the cross‐section measurement
procedures described in 1372HFigure 5.13.) The hole size and web location dimensions, based
on the average of three independent measurements, are provided for each specimen in
1373HTable 5.8.
183
Table 5.8 Measured slotted hole dimensions and locations X W1 W2 L hole h hole X W1 W2 L hole h hole
in. in. in. in. in. in. in. in. in. in.362-1-24-H L/2 0.946 1.141 4.003 1.492362-2-24-H L/2 1.146 0.967 4.000 1.502362-3-24-H L/2 0.935 1.114 4.005 1.493362-1-48-H (L-24)/2 1.252 0.974 3.999 1.500 (L+24)/2 1.198 0.952 4.001 1.494362-2-48-H (L-24)/2 1.126 1.016 4.001 1.496 (L+24)/2 1.171 0.973 4.003 1.494362-3-48-H (L-24)/2 0.982 1.112 4.000 1.493 (L+24)/2 0.967 1.133 4.003 1.491600-1-24-H L/2 2.147 2.361 4.002 1.498600-2-24-H L/2 2.365 2.155 4.001 1.491600-3-24-H L/2 2.347 2.166 4.001 1.493600-1-48-H (L-24)/2 2.161 2.375 4.002 1.494 (L+24)/2 2.162 2.383 3.998 1.497600-2-48-H (L-24)/2 2.166 2.351 4.001 1.499 (L+24)/2 2.176 2.360 4.002 1.498600-3-48-H (L-24)/2 2.371 2.162 3.999 1.497 (L+24)/2 2.365 2.156 4.003 1.494
Specimen
2.4.6 147BWeb imperfections
Variations in the specimen webs were measured to provide a basis for the local
buckling initial imperfection magnitudes in the specimen nonlinear finite element
models constructed in Section 1374H7.2. The measurement setup shown in 1375HFigure 5.20 uses a
dial gauge with a precision of 0.001 inches mounted to a laboratory stand in contact with
a flat steel table. The specimen was supported horizontally at both ends by a matching
pair of steel bars that were ground flat and parallel. The bars were also in contact with
the steel table, ensuring that the specimen and the dial gauge were in the same
horizontal reference plane. Each specimen web was marked with a grid of measurement
points shown in 1376HFigure 5.21. The stand and dial gauge were shifted from grid point to
grid point and elevation measurements were recorded. The variations from the average
elevation of the specimen web, based on an average of two measurements per grid
point, are provided for each specimen in 1377HTable 5.9.
184
Figure 5.20 A dial gauge and precision stand are used to measure initial web imperfections
North
X
1.2 inches (362 specimens)2.3 inches (600 specimens)
CL Web (typ.)
West Center East
a
a
Section a-a
+ variation
Plan view(short and intermediate length web grid layouts)
6 in. (typ.)
Figure 5.21 Web imperfection measurement grid and coordinate system
185
Table 5.9 Initial web imperfections (deviations from the average elevation of the web)
Specimen X Distance in. 0 6 12 18 24 30 36 42 48West in. 0.013 -0.007 -0.011 -0.004 0.015Center in. 0.022 -0.005 -0.022 -0.013 0.015East in. 0.013 -0.007 -0.013 -0.004 0.014West in. 0.019 -0.006 -0.010 -0.006 0.015Center in. 0.015 -0.014 -0.020 -0.007 0.024East in. 0.015 -0.009 -0.015 -0.008 0.014West in. 0.016 -0.004 -0.010 -0.003 0.015Center in. 0.017 -0.015 -0.023 -0.003 0.025East in. 0.016 -0.010 -0.016 -0.008 0.014West in. 0.006 -0.008 -0.014 -0.001 0.016Center in. 0.016 -0.010 Hole -0.009 0.009East in. 0.009 -0.008 -0.013 -0.001 0.015West in. 0.007 -0.009 -0.020 -0.003 0.014Center in. 0.014 -0.014 Hole -0.007 0.010East in. 0.025 -0.001 -0.017 -0.009 0.014West in. 0.016 -0.009 -0.020 -0.010 0.016Center in. 0.021 -0.009 Hole -0.015 0.015East in. 0.017 -0.002 -0.015 -0.002 0.015West in. 0.003 -0.010 -0.014 -0.011 -0.009 -0.005 -0.005 -0.002 0.018Center in. 0.009 0.004 -0.006 -0.006 -0.005 -0.007 -0.005 -0.008 0.013East in. 0.015 0.019 0.010 0.005 0.004 -0.001 -0.002 -0.005 0.004West in. -0.008 -0.023 -0.021 -0.013 -0.002 0.006 0.015 0.021 0.023Center in. -0.004 -0.016 -0.021 -0.015 0.000 0.006 0.011 0.012 -0.002East in. -0.005 -0.008 -0.011 -0.009 0.004 0.010 0.013 0.016 0.012West in. 0.006 -0.002 0.003 0.005 0.003 0.002 0.000 0.011 0.007Center in. -0.001 -0.003 0.002 0.001 -0.002 -0.004 -0.011 -0.001 0.004East in. 0.010 0.007 0.006 0.003 -0.001 -0.007 -0.013 -0.011 -0.001West in. -0.006 -0.003 -0.003 0.002 0.007 0.007 0.007 0.010 0.000Center in. -0.015 0.003 Hole 0.001 0.006 -0.002 Hole -0.009 -0.009East in. 0.009 0.021 0.016 0.010 0.007 -0.004 -0.016 -0.022 -0.015West in. 0.013 -0.007 -0.003 -0.001 0.000 -0.004 -0.009 -0.002 0.016Center in. 0.011 -0.011 Hole -0.006 -0.003 -0.010 Hole -0.004 0.022East in. 0.010 -0.002 -0.004 -0.001 -0.003 -0.004 -0.012 -0.006 0.012West in. 0.013 -0.007 -0.012 -0.006 -0.003 0.003 -0.003 -0.004 0.017Center in. 0.019 -0.005 Hole -0.010 -0.002 -0.003 Hole -0.008 0.015East in. 0.014 -0.002 -0.010 -0.006 0.001 0.001 -0.007 -0.002 0.012West in. 0.016 -0.012 -0.029 -0.024 0.013Center in. 0.055 0.005 -0.027 -0.023 0.027East in. 0.016 -0.003 -0.014 -0.014 0.009West in. 0.013 -0.019 -0.033 -0.027 0.009Center in. 0.061 0.004 -0.030 -0.024 0.031East in. 0.021 0.002 -0.010 -0.011 0.014West in. 0.007 -0.016 -0.018 -0.003 0.010Center in. 0.034 -0.023 -0.029 0.006 0.057East in. 0.017 -0.021 -0.028 -0.012 0.018West in. 0.005 -0.015 -0.031 -0.019 0.011Center in. 0.052 0.003 Hole -0.017 0.021East in. 0.020 -0.003 -0.018 -0.012 0.006West in. 0.009 -0.014 -0.024 -0.011 0.012Center in. 0.020 -0.018 Hole -0.001 0.051East in. 0.014 -0.015 -0.027 -0.013 0.016West in. 0.006 -0.001 -0.003 -0.006 0.014Center in. 0.007 -0.009 Hole -0.002 0.040East in. 0.007 -0.014 -0.022 -0.016 0.004West in. 0.023 -0.003 -0.018 -0.026 -0.026 -0.020 -0.019 -0.014 0.016Center in. 0.060 0.016 -0.010 -0.018 -0.010 -0.006 -0.009 -0.005 0.030East in. 0.024 0.006 -0.002 -0.008 -0.001 0.003 0.004 0.001 0.011West in. 0.019 -0.004 -0.016 -0.016 -0.020 -0.025 -0.024 -0.014 0.011Center in. 0.060 0.012 -0.005 -0.004 -0.007 -0.009 -0.009 -0.006 0.023East in. 0.014 -0.001 0.002 0.005 0.000 -0.004 0.003 0.002 0.008West in. 0.026 -0.003 -0.021 -0.021 -0.014 -0.011 -0.015 -0.015 0.013Center in. 0.055 0.013 -0.012 -0.011 0.003 0.003 -0.008 -0.006 0.031East in. 0.013 -0.003 -0.010 -0.010 0.002 0.008 -0.002 -0.002 0.008West in. 0.014 -0.004 -0.026 -0.026 -0.026 -0.024 -0.025 -0.007 0.009Center in. 0.059 0.012 Hole -0.007 0.002 -0.009 Hole -0.002 0.024East in. 0.009 0.003 0.000 0.002 0.004 0.002 0.001 0.004 0.002West in. 0.032 0.002 -0.020 -0.028 -0.022 -0.012 -0.019 -0.022 0.013Center in. -0.033 0.023 Hole -0.001 0.006 0.005 Hole 0.002 0.025East in. 0.011 0.008 0.004 0.004 0.004 0.007 0.000 -0.004 0.004West in. 0.017 0.012 0.012 0.010 0.000 0.007 0.009 0.011 0.020Center in. 0.028 0.003 Hole -0.004 -0.022 -0.014 Hole 0.018 0.046East in. 0.018 -0.010 -0.023 -0.032 -0.048 -0.040 -0.045 0.002 0.021
362-1-24-NH
362-2-24-NH
362-3-24-NH
362-1-24-H
362-2-48-H
362-3-48-H
362-2-24-H
362-3-24-H
362-1-48-NH
362-2-48-NH
600-3-48-H
600-2-24-H
600-3-24-H
600-1-48-NH
600-2-48-NH
Local Variations in Web
600-3-48-NH
600-1-48-H
600-2-48-H
600-1-24-NH
600-2-24-NH
600-3-24-NH
600-1-24-H
362-3-48-NH
362-1-48-H
186
2.4.7 148BSpecimen orientation in the testing machine
When placing the specimen in the testing machine, the southern end of the
specimen was oriented at the bottom platen such that the center of the compressive force
was applied through the gross centroid of the C‐section. The centerline of the web is
positioned in line with the centerline of the bottom platen and offset towards the back of
the testing machine as described in 1378HFigure 5.22. The centroid locations were calculated
using the centerline dimensions of a nominal SSMA 362S162‐33 and 600S162‐33 cross
section.
Plan View(Bottom Platen)
CL Platen and Column Web
0.380 in. (600S162-33)0.502 in. (362S162-33)
CL Platen
Location of interior web edge
Center of platen, center of load, centroid of Ceechannel
FRONT OF MTS MACHINE
Column specimen
Figure 5.22 Column specimen alignment schematic
187
The actual cross section and thickness measurements produced centroid offsets
slightly different from the nominal offsets considered in the column tests. The
difference between the nominal and measured offsets, defined here as ΔCG, are
provided in 1379HTable 5.10. ΔCG produces end moments in the specimens that are several
orders of magnitude smaller than the applied loads in this study. For example, the end
moments created by a ΔCG of 0.059 inches for specimen 600‐3‐24‐NH are calculated as
2.0 x 10‐6 kip∙inches at peak load (Ptest=12.24 kips) using the structural analysis program
MASTAN (Ziemian and McGuire 2005). The assumed MASTAN structural system in
1380HFigure 5.23 demonstrates that relatively stiff compression platens and fixed‐fixed end
conditions effectively eliminate end moments from small load eccentricities.
Table 5.10 Specimen gross centroid and offset from applied load during tests
ΔCSxcg tz xcg - tz used in testsin. in. in. in. in.
362-1-24-NH 0.482 0.038 0.463 0.502 0.039362-2-24-NH 0.471 0.038 0.452 0.502 0.050362-3-24-NH 0.504 0.038 0.485 0.502 0.017362-1-24-H 0.511 0.042 0.489 0.502 0.013362-2-24-H 0.490 0.042 0.469 0.502 0.033362-3-24-H 0.524 0.042 0.503 0.502 -0.001362-1-48-NH 0.475 0.041 0.454 0.502 0.048362-2-48-NH 0.468 0.042 0.447 0.502 0.055362-3-48-NH 0.475 0.040 0.455 0.502 0.047362-1-48-H 0.470 0.041 0.449 0.502 0.053362-2-48-H 0.470 0.042 0.449 0.502 0.053362-3-48-H 0.486 0.040 0.466 0.502 0.036600-1-24-NH 0.354 0.047 0.330 0.380 0.050600-2-24-NH 0.347 0.047 0.323 0.380 0.057600-3-24-NH 0.344 0.047 0.321 0.380 0.059600-1-24-H 0.363 0.046 0.340 0.380 0.040600-2-24-H 0.368 0.047 0.344 0.380 0.036600-3-24-H 0.361 0.046 0.338 0.380 0.042600-1-48-NH 0.362 0.046 0.339 0.380 0.041600-2-48-NH 0.355 0.045 0.333 0.380 0.047600-3-48-NH 0.353 0.045 0.330 0.380 0.050600-1-48-H 0.362 0.045 0.340 0.380 0.040600-2-48-H 0.352 0.046 0.329 0.380 0.051600-3-48-H 0.356 0.046 0.333 0.380 0.047
tz sheet thickness with zinc coatingΔCS difference measured and as tested centroid offsets
SpecimenSpecimen
Measurements Centroid Shift
188
Actuator load
Column specimen
(Centroid shown)
ΔCG
Platen (typ.)
Horizontal translation and rotational DOF restrained
CL Applied Load
Stiff platen Flexible platen
Structural System Moment Diagrams
No moment in column for stiff platen even with load offset
All translation and rotational DOF restrained
EI
1000EI
EI
0.1EI
1000EI 0.1EI
EI flexural rigidity
Figure 5.23 Influence of platen bending stiffness on end moments for a fixed‐fixed eccentric column
Once the specimen is aligned on the bottom platen, 500 lbs of compressive force
was applied to the column and weak‐axis out of straightness measurements were taken.
The distance from the front of the top and bottom platens to the interior web edge is
denoted as Stop and Sbottom in 1381HFigure 5.24. Stop and Sbottom are obtained as the average of three
independent measurements with digital calipers as shown in 1382HFigure 5.25 and then
corrected for a systematic platen offset (see 1383HFigure 5.24) and the initial web
imperfections in 1384HTable 5.9. The initial out‐of‐straightness ΔS provided in 1385HTable 5.11 is
189
calculated from Stop and Sbottom and implemented as an initial geometric imperfection into
the nonlinear finite element models in 1386H7.2.
a a
Sbottom
Stop
Stop
Section a-a
Column Specimen(orientation exaggerated)
Front of MTS Machine
Side View(Looking west)
CL Platen
CL Platen
CL Load
ΔS (negative magnitude shown)
Platen Offset=0.084 in.
Figure 5.24 Column specimen weak axis out‐of‐straightness schematic
190
Figure 5.25 Digital calipers are used to measure the distance from the column web to platen edge
Table 5.11 Summary of out‐of‐straightness calculations Specimen Sbottom Stop Platen Offset Sbottom Correction Sbottom Correction Stop ΔS
As measured
As measured
Top platen edge is offset from bottom
platen Edge
Corrected for top platen
offset
Web Imperfection @
X=0
Corrected for Web
Imperfection @ X=0
Web Imperfection
@ X=L
Corrected for Web
Imperfection @ X=L
Initial out of straightness
in. in. in. in. in. in. in. in. in.362-1-24-NH 6.507 6.622 0.084 6.591 0.015 6.577 0.022 6.600 -0.024362-2-24-NH 6.523 6.612 0.084 6.607 0.015 6.593 0.024 6.588 0.004362-3-24-NH 6.531 6.585 0.084 6.615 0.017 6.598 0.025 6.560 0.038362-1-24-H 6.524 6.613 0.084 6.608 0.016 6.592 0.009 6.604 -0.012362-2-24-H 6.532 6.578 0.084 6.616 0.014 6.602 0.010 6.568 0.034362-3-24-H 6.529 6.629 0.084 6.613 0.021 6.592 0.015 6.615 -0.023362-1-48-NH 6.352 6.393 0.084 6.436 0.009 6.427 0.013 6.380 0.047362-2-48-NH 6.535 6.649 0.084 6.619 -0.004 6.623 -0.002 6.651 -0.028362-3-48-NH 6.537 6.614 0.084 6.621 -0.001 6.622 0.004 6.610 0.012362-1-48-H 6.530 6.554 0.084 6.614 -0.015 6.629 -0.009 6.563 0.066362-2-48-H 6.534 6.617 0.084 6.618 0.011 6.607 0.022 6.594 0.013362-3-48-H 6.532 6.616 0.084 6.616 0.019 6.598 0.015 6.601 -0.003600-1-24-NH 6.352 6.472 0.084 6.436 0.055 6.381 0.027 6.444 -0.063600-2-24-NH 6.365 6.560 0.084 6.449 0.061 6.388 0.031 6.529 -0.141600-3-24-NH 6.451 6.494 0.084 6.535 0.034 6.501 0.057 6.437 0.063600-1-24-H 6.356 6.486 0.084 6.440 0.052 6.388 0.021 6.466 -0.078600-2-24-H 6.360 6.399 0.084 6.444 0.020 6.424 0.051 6.348 0.076600-3-24-H 6.355 6.403 0.084 6.439 0.007 6.432 0.040 6.363 0.069600-1-48-NH 6.346 6.436 0.084 6.430 0.060 6.370 0.030 6.406 -0.036600-2-48-NH 6.354 6.488 0.084 6.438 0.060 6.377 0.023 6.465 -0.087600-3-48-NH 6.354 6.463 0.084 6.438 0.055 6.383 0.031 6.432 -0.049600-1-48-H 6.311 6.458 0.084 6.395 0.059 6.336 0.024 6.433 -0.098600-2-48-H 6.352 6.422 0.084 6.436 -0.033 6.469 0.025 6.396 0.072600-3-48-H 6.348 6.430 0.084 6.432 0.028 6.404 0.046 6.384 0.020
191
2.5 80BMaterials testing
Tensile coupon tests were performed to obtain the steel stress‐strain curve and
yield stress for the web, west flange, and east flange of each specimen in this study. The
tests were conducted in accordance with ASTM specification E 8M‐04, “Standard Test
Methods for Tension Testing of Metallic Materials (Metric)” (ASTM 2004).
2.5.1 149BTensile coupon preparation Tensile coupons were always obtained from the same 8 ft. structural stud which
produced the column specimen. Flat portions of the web and flanges were first rough
cut with a metal band saw as shown in 1387HFigure 5.26, and then finished to the dimensions
in 1388HFigure 5.27 with a CNC milling machine. The special jig in 1389HFigure 5.27 allowed for
three tensile coupons to be produced at once. The tensile coupons were stripped of their
zinc coating (see Section 1390H5.2.4.3 for procedure) and then measured within the gauge
length for bare metal thickness, t, and minimum width, wmin. The minimum width is
determined by taking the minimum of five independent measurements within the gauge
length of the specimen with digital calipers.
192
Figure 5.26 Tensile coupons are first rough cut with a metal ban saw
0.79 in.
1.97 in.
1.97 in.
3.18 in. 1.97 in.
0.38 in.0.38 in.
R=0.55 in.
0.492 in. *
gauge length
*nominal, actual dimension will vary slightly
Figure 5.27 Tensile coupon dimensions as entered in the CNC milling machine computer
193
Figure 5.28 A custom jig allows three tensile coupons to be milled at once in the CNC machine
2.5.2 150BTensile test setup A screw‐driven ATS 900 testing machine with a maximum capacity of 10 kips
was used to apply the tensile load. Tensile coupons were positioned in the machine
with friction grips as shown in 1391HFigure 5.29. A bubble level with a short, straight edge
was used to ensure that each specimen was aligned vertically between the grips. An
MTS 634.11D‐54 extensometer measured engineering strain and an MTS load cell
measured force on the specimen. The extensometer was placed at the vertical midlength
of the specimen, centered within the gauge length. The raw voltage data from the
extensometer and load cell were sent to a PC containing a National Instruments data
acquisition card. The voltage data was converted to tensile force and engineering strain
using the conversion factors provided in 1392HTable 5.12. The data was plotted on the screen
and recorded to a file with a custom LabVIEW program (Labview 2005).
194
Figure 5.29 ATS machine used to test tensile coupons
Table 5.12 Voltage conversion factors for tensile coupon testing
Measurement Source ConversionTensile Force MTS Load Cell 1 Volt = 1000 lbf
Engineering Strain MTS Extensometer 1 Volt = 3.96x10-5 strain (in./in.)
Tensile Coupon Testing
2.5.3 151BTensile test results
Two distinct steel stress‐strain curves were observed in this study. Tensile
coupons from the 362S162‐33 structural studs demonstrate gradual yielding behavior,
while the tensile coupons from the 600S162‐33 studs demonstrated a sharp yielding
plateau. The yield stress, Fy, for the gradually yielding specimens was determined with
the 0.2% strain offset method. The stress‐strain curve for specimen 362‐3‐48‐NH (East
Flange) demonstrates the offset method in 1393HFigure 5.30. The yield stress for the sharply
yielding specimens was determined by averaging the stresses in the yield plateau.
ASTM does not provide specific guidelines on how to average the plateau stresses. For
this autographic method, the averaging range is determined by using two strain offset
lines, one at 0.4% strain offset and the other at 0.8% offset as shown for specimen 600‐24‐
195
NH (West Flange) in 1394HFigure 5.31. The steel modulus of elasticity, E, was assumed as
29500 ksi for all specimens when determining the yield stress. The tensile coupon yield
stresses and cross section dimensions are summarized in 1395HTable 5.13. The mean and
standard deviation for all 362S162‐33 and 600S162‐33 tensile coupons tested are
provided in 1396HTable 5.14.
0 0.2 0.4 0.6 0.8 10
0.2
0.4
0.6
0.8
1
0 0.05 0.1 0.15 0.2 0.250
10
20
30
40
50
60
70
80
90
100
Engineering Strain,(in./in.)
Axi
al T
ensi
le S
tress
(ksi
)
YIELD STRESS (0.2% offset)=60.1 ksi
0.2% strain offset line (slope=29500 ksi)
Figure 5.30 Gradually yielding stress‐strain curve with 0.2% strain offset method
0 0.2 0.4 0.6 0.8 10
0.2
0.4
0.6
0.8
1
0 0.05 0.1 0.15 0.2 0.250
10
20
30
40
50
60
70
80
90
100
Engineering Strain (in./in.)
Axi
al T
ensi
le S
tress
(ksi
)
YIELD STRESS (Autographic Method)=59.7 ksi
0.4% strain offset line (slope=29500 ksi)
0.8% strain offset line (slope=29500 ksi)
Figure 5.31 Sharp‐yielding stress strain curve using an autographic method for determining Fy
Table 5.13 Summary of column specimen steel yield stress
196
tbase,w wmin Fy tbase,f1 wmin Fy tbase,f2 wmin Fy
in. in. ksi in. in. ksi in. in. ksi362-1-24-NH362-2-24-NH362-3-24-NH362-1-24-H 0.0390 0.4945 55.9 0.0391 0.4963 59.3 0.0391 0.4968 58.5362-2-24-H 0.0368 0.4886 52.9 0.0390 0.4950 58.8 0.0391 0.4945 59.5362-3-24-H 0.0394 0.4945 55.6 0.0394 0.4927 N/C 0.0394 0.4947 56.4362-1-48-NH 0.0392 0.4985 59.4 0.0393 0.4965 59.7 0.0392 0.4975 59.9362-2-48-NH 0.0393 0.4990 59.2 0.0394 0.4975 59.3 0.0393 0.4970 59.2362-3-48-NH 0.0389 0.4930 58.0 0.0391 0.5000 58.9 0.0390 0.4930 60.1362-1-48-H 0.0391 0.4998 59.5 0.0393 0.4985 58.2 0.0394 0.4991 58.1362-2-48-H 0.0390 0.4992 58.8 0.0391 0.4961 60.6 0.0391 0.4975 59.8362-3-48-H 0.0401 0.4990 57.8 0.0400 0.4957 58.0 0.0397 0.4978 59.1600-1-24-NH600-2-24-NH600-3-24-NH600-1-24-H 0.0414 0.4899 61.9 0.0422 0.4940 63.6 0.0428 0.4964 60.3600-2-24-H 0.0427 0.4964 57.8 0.0384 0.4874 55.6 0.0424 0.4938 61.8600-3-24-H 0.0429 0.4966 59.7 0.0431 0.4954 58.0 0.0430 0.4960 62.6600-1-48-NH 0.0434 0.4985 58.7 0.0436 0.4955 62.3 0.0434 0.4965 59.3600-2-48-NH 0.0435 0.4985 N/C 0.0430 0.4970 63.4 0.0430 0.4970 63.3600-3-48-NH 0.0436 0.4995 60.4 0.0432 0.4955 N/C 0.0433 0.4965 61.9600-1-48-H 0.0429 0.4970 60.3 0.0426 0.4980 63.0 0.0429 0.4970 60.8600-2-48-H 0.0429 0.4994 61.8 0.0428 0.4962 62.1 0.0431 0.4977 62.2600-3-48-H 0.0430 0.4992 60.7 0.0434 0.4961 59.7 0.0430 0.4977 64.0NOTE: N/C Tests results were not obtained
0.5000 55.90.4950 59.7
57.40.4975 54.7
SpecimenWeb East FlangeWest Flange
0.0368 0.03720.0415 0.4955
0.0438 0.04380.0432
0.4985 53.3
0.4950 60.6
Table 5.14 Column specimen steel yield stress statistics
mean STDVksi ksi
362S162-33 58.1 2.0600S162-33 61.0 2.0
yield stress, FyStud Type
5.3 36B Elastic buckling calculations
Elastic buckling provides a means to categorize and potentially better understand
the load‐deformation response and ultimate strength of the thin‐walled columns in this
study. The local, distortional, and global elastic buckling modes and their associated
critical elastic buckling loads (Pcrl, Pcrd, Pcre) are presented here for each specimen.
Calculations are performed with a shell finite element eigenbuckling analysis as
opposed to an analysis using FSM (Schafer and Ádàny 2006) to capture the influence of
the slotted web holes and the tested (fixed‐fixed) boundary conditions.
197
3.1 81BFinite element modeling assumptions
Eigenbuckling analysis in ABAQUS is performed for the 24 column specimens
(ABAQUS 2007a). All columns are modeled with S9R5 reduced integration nine‐node
thin shell elements. Cold‐formed steel material properties are assumed as E=29500 ksi
and ν=0.30. The centerline C‐section dimensions input into ABAQUS are calculated
using the out‐to‐out dimensions and flange and lip angles at the mid‐height of each
column specimen as provided in 1397HTable 5.3 and 1398HTable 5.4. Each column specimen is
loaded with a set of consistent nodal loads in ABAQUS to simulate a constant pressure
across the bearing edge of the specimen. The nodes on the loaded column face are
coupled together in the direction of loading with an ABAQUS “pinned” rigid body
constraint (see 1399HFigure 4.12).
3.2 82BElastic buckling results
3.2.1 152BBuckled shapes / eigenmodes
The first (lowest buckling load) local (L) and distortional (D) buckled shapes for
specimens with and without slotted holes are compared in 1400HFigure 5.32 and 1401HFigure 5.33.
The L and D modes for each specimen were identified visually by manually searching
through the elastic buckling modes produced in the eigenbucking analysis. The nominal
cross‐section half‐wavelengths in 1402HTable 5.1 were compared to the half‐wavelengths in
the finite element model to assist in the categorization. The local and distortional modes
that most resembled the FSM results for L and D modes were selected. This method of
modal identification is neither exact nor ideal, especially when both local and
198
distortional buckling are present in the same eigenmode. Formal modal identification
has recently been developed in the context of the finite strip method (Schafer and Ádàny
2006) and future work is ongoing to extend this method to finite element analyses and to
problems such as the ones encountered here.
Local Buckling Distortional Buckling
Hole terminates web local buckling
Holes cause mixed distortional-local mode
Local Buckling (L) Distortional Buckling (D)
Hole terminates web local buckling
(a) (b)
Figure 5.32 (a) Local and distortional elastic buckled mode shapes for (a) short (L=48 in.) 362S162‐33 specimens and (b) intermediate length (L=48 in.) 362S162‐33 specimens.
Local Buckling Distortional Buckling
Holes change number of half-waves from 8 (NH) to 12 (H)
Holes cause mixed distortional – local mode
Local Buckling (L) Distortional Buckling (D)
Hole changes number of half-waves from 5 (NH) to 6 (H)
(a) (b)
Figure 5.33 Local and distortional elastic buckled mode shapes for (a) short (L=48 in.) 600S162‐33 specimens and (b) intermediate length (L=48 in.) 600S162‐33 specimens.
3.2.2 153BBuckling loads / eigenvalues
The primary goal of this research program is to extend the Direct Strength Method to
cold‐formed steel structural members with holes. The Direct Strength Method (DSM), a
design method for cold‐formed steel structural members, predicts column ultimate
strength by predicting the column failure mode and ultimate strength through
199
knowledge of the local (L), distortional (D), or global (G) elastic buckling modes. This
connection is made using the critical elastic buckling load, Pcr, and the slenderness,
defined with the ratio of column squash load Pyg to Pcr for the L, D, and G modes. 1403HTable
5.15 summarizes Pcr and Pyg for the specimens evaluated in this study. The squash load
Pyg is calculated with the gross cross‐sectional area, and Pcr includes the effects of the
holes and the tested (fixed‐fixed) boundary conditions. (Note, the implications of using
Pyg as opposed to Py,net at the net section are discussed in 1404HChapter 8.)
The influence of holes on Pcr is of interest in the context of DSM because elastic
buckling loads and slenderness are used to predict ultimate strength. To isolate the
influence of holes on Pcr, additional eigenbuckling analyses of the specimens with holes
(specimens labeled with an H) were performed, but with the holes removed (the
boundary and loading conditions were not modified and the mesh used in the models
was identical except for the removed elements at the hole location). The comparison of
Pcr for specimens with holes (H) and then with holes removed (noH) is also summarized
in 1405HTable 5.15.
200
Table 5.15 Critical elastic buckling loads, influence of holes on elastic buckling
Pyg Pcre Pcrl Pcrd
kips kips kips kips
362-1-24-NH 15.5 109.4 4.9 10.6362-2-24-NH 15.6 112.5 4.8 10.2362-3-24-NH 15.7 112.2 5.0 10.7362-1-24-H 16.4 119.3 5.9 13.5 0.98 1.03 1.12362-2-24-H 15.7 112.8 5.4 12.4 0.98 1.02 1.13362-3-24-H 16.4 130.6 5.7 12.9 0.99 1.02 1.12
362-1-48-NH 16.9 30.5 5.2 9.7362-2-48-NH 16.7 29.5 5.2 9.6362-3-48-NH 16.6 29.6 5.1 9.5362-1-48-H 16.6 30.0 5.3 9.4 0.94 1.03 0.98362-2-48-H 16.8 29.7 5.2 9.3 0.94 1.03 0.98362-3-48-H 16.8 36.2 5.7 9.6 0.95 1.03 0.98
600-1-24-NH 24.7 244.5 3.4 6.8600-2-24-NH 24.5 234.9 3.4 6.7600-3-24-NH 24.5 218.4 3.4 6.6600-1-24-H 25.0 239.3 3.3 7.0 1.01 1.02 1.09600-2-24-H 23.1 238.4 3.2 6.7 1.01 1.01 1.08600-3-24-H 24.7 242.6 3.5 7.3 1.02 1.01 1.08
600-1-48-NH 25.1 61.8 3.5 5.2600-2-48-NH 26.2 59.6 3.4 5.7600-3-48-NH 25.4 60.2 3.4 5.7600-1-48-H 25.2 56.3 3.4 5.1 0.87 1.02 1.02600-2-48-H 25.5 53.0 3.4 5.0 0.87 1.02 1.02600-3-48-H 25.6 55.8 3.4 5.0 0.86 1.02 1.02
* For specimens with holes (H), the holes are removed and elastic buckling calculated (noH). The hole (H) and no hole (noH) finite element models are otherwise identical, isolating the influence of the holes.
Specimen Name
ELASTIC BUCKLING HOLE INFLUENCE*
Pcre/PcrenoH Pcrl/Pcrl
noHPcrd/Pcrd
noH
N/A
N/A
N/A
N/A
3.2.3 154BModal interaction at ultimate strength
An additional reason for the selection of these specimen cross‐sections, at these
lengths, beyond the reasons discussed in Section 1406H5.2.1, is that the specimens provide
much needed experimental data on cross‐sections with potential modal interaction at
ultimate strength both with and without holes. Typically modal interaction is
understood to be a concern when the elastic buckling loads of multiple modes are at or
near the same value, and the ratio of any two elastic bucking loads (e.g., Pcrl/Pcrd) is
considered a useful parameter for study. However, for modes with different post‐
buckling strength and where material yielding is considered, a more pressing concern
may be the situation when both failure modes predict similar capacities. Which mode
does the column fail in if the predicted capacity in local (Pnl) and distortional (Pnd) are at
or near the same level? What impact does a hole have on the failure mode that is
201
triggered? In the specimens selected here, the ratio of Pcrl/Pcrd varies from a min of 0.44 to
a max of 0.68, but is never near 1.0. Therefore, by this traditional measure no meaningful
interaction would be anticipated. However, if the DSM methodology is used to predict
the capacities, as illustrated in 1407HFigure 5.34, the predictions for the ratio of the two limit
states Pnl/ Pnd ranges from a min of 0.86 to a max of 0.90 in the 362S162‐33 short columns
and from a min of 1.0 to a max of 1.05 in the 600S162‐33 short columns (the ratios are
similar for the long column specimens). Thus, these cross‐sections provide a means to
examine the potential for local‐distortional modal interaction at ultimate strength, and
offer valuable data for determining any necessary modification to the DSM
methodology when holes are present.
0 0.5 1 1.5 2 2.5 30
0.2
0.4
0.6
0.8
1
1.2
slenderness, λl or λd
Pn/P
yg
Local DSMDistortional DSM
D
D
L
L
362S162-33 short columns
600S162-33 short columns
Local(L)-Distortional(D) interaction is expected since predicted strengths (Pn) are of similar magnitudes
Figure 5.34 Local (L) and distortional (D) DSM strength predictions are similar in magnitude for both
362S162‐33 and 600S162‐33 cross‐sections, indicating that L‐D modal interaction will occur during the tested response of the columns.
202
3.3 83BDiscussion of elastic buckling results
3.3.1 155BLocal buckling
Boundary conditions have little influence on the local buckling mode shapes
(compared with FSM L modes), but the presence of the slotted web holes can change the
shape, half‐wavelength, and buckling load of the first (lowest) local buckling mode
observed. In the 362S162‐33 specimens the web holes terminate local buckling in the
vicinity of the holes, see 1408HFigure 5.32. In the 600S162‐33 specimens the web holes cause
an increased number of half‐waves along the length to occur in the lowest local mode,
see 1409HFigure 5.33. The presence of holes causes a slight increase in Pcrl (see 1410HTable 5.15)
which is consistent with the increased number of observed local buckling half‐waves.
The more extensive elastic buckling studies 1411HChapter 3 and 1412HChapter 4 demonstrate that a
hole can increase or decrease the number of buckled half‐waves (and the critical elastic
buckling load) of rectangular plates and cold‐formed steel structural studs.
3.3.2 156BDistortional buckling
Boundary conditions and the presence of holes have an influence on the observed
distortional buckling mode shapes (compared with FSM D modes) and buckling loads.
The boundary conditions (fixed‐fixed) allow a smaller number of half‐waves to form
than predicted using the simply supported FSM D modes of 1413HTable 5.1. For example,
observe the restrained shape of the buckled distortional half‐wave near the member
ends in 1414HFigure 5.32a. In longer specimens (see 1415HFigure 5.32b and 1416HFigure 5.33b), the
influence of the boundary conditions lessens and the half‐wavelength of distortional
203
buckling at mid‐height approaches that of the FSM D mode of 1417HTable 5.1. (Section 1418H4.2.6.2
explores the influence of fixed‐fixed boundary conditions on Pcrd using the column
experiment database.) The presence of the web holes complicates the predicted D
modes, see 1419HFigure 5.32 and 1420HFigure 5.33. Local buckling now appears within the D mode
itself. The half‐wavelength of these interacting L modes is significantly shorter than the
lowest L modes observed. Further, and rather unintuitively, the buckling load, Pcrd,
actually increases with the presence of holes in the short column specimens (as much as
13%). However, this increase is lost at the longer specimen length where the maximum
change in the buckling load is +/‐ 2%. This result suggests that in the shorter specimens
the removal of the material most susceptible to out‐of‐plane bending, at the mid‐depth
of the web, actually serves to stiffen the column (a localized increase in the transverse
bending stiffness of plates with holes has been observed, see 1421HFigure 4.30). This influence
does not persist in the longer specimens suggesting that the increased stiffness is only
relevant when the D mode is at a restrained half‐wavelength. Thus, if the D mode is free
to form (over a long enough unbraced length) the holes do not increase the elastic
buckling load.
3.3.3 157BGlobal buckling
The global (Euler) buckled shapes for the intermediate 362S162‐33 and 600S162‐33
columns in 1422HFigure 5.35 occur as flexural‐torsional buckling, although local and
distortional deformation are both present in the mode shape for specimens with and
without holes, which is an unexpected result. The interaction between the global, local,
204
and distortional modes makes the identification of the global mode difficult. The Euler
buckling load and mode shape predicted with classical methods (in CUTWP), which do
not allow cross‐section distortion and ignore holes, were used to determine the range of
buckling loads (eigenvalues) to be visually searched. The reported modes in 1423HFigure 5.35
are the ones closest to the expected buckling load exhibiting significant global
deformations. Additional eigenbuckling analyses of the 362S162‐33 and 600S162‐33
cross‐sections were performed at a longer column length (8 ft.) and these analyses show
no local or distortional interaction with the global modes. Therefore, the observed
interaction is length dependent and not a fundamental feature of global buckling in
these cross‐sections. An alternative hypothesis for the “unusual” mode shapes in 1424HFigure
5.35 is that several buckling mode shapes exist near the global critical elastic buckling
load, which causes the eigensolver to misreport the global mode as a linear combination
of buckled shapes.
As for the global buckling loads, the slotted holes have a small influence on the
global buckling load for the intermediate length 362S162‐33 specimens, reducing Pcre by a
maximum of 6%. However, Pcre for the intermediate length 600S162‐33 columns
decreases by a maximum of 14% with the presence of the two slotted holes, which is an
unexpected result attributed to the local and distortional modes mixing with global
buckling (i.e., 1425HFigure 5.35). Additional research work is ongoing to determine under
what conditions holes influence the global critical elastic buckling load.
205
362-1-48-NH
600-1-48-H362-1-48-H
600-1-48-NH
CUTWP predictions using classical stability theory
Finite element eigenbucklinganalyses predict global buckling interacting with local and distortional buckling
Global (Euler) Buckling
Figure 5.35 Comparison of global mode shapes for intermediate length 362S162‐33 and 600S162‐33 specimens.
5.4 37BExperiment results
4.1 84BUltimate strength
The peak tested compressive load for all column specimens and an average peak
load for each test group are provided in 1426HTable 5.16. The slotted holes are shown to have
only a small influence on compressive strength in this study, with the largest reduction
being 2.7% for the 362S162‐33 short columns.
206
Table 5.16 Specimen ultimate strength results Ptest Mean Std. Dev.kips kips kips
362-1-24-NH 10.48362-2-24-NH 10.51362-3-24-NH 10.15362-1-24-H 10.00362-2-24-H 10.38362-3-24-H 9.94362-1-48-NH 9.09362-2-48-NH 9.49362-3-48-NH 9.48362-1-48-H 8.95362-2-48-H 9.18362-3-48-H 9.37600-1-24-NH 11.93600-2-24-NH 11.95600-3-24-NH 12.24600-1-24-H 12.14600-2-24-H 11.62600-3-24-H 11.79600-1-48-NH 11.15600-2-48-NH 11.44600-3-48-NH 11.29600-1-48-H 11.16600-2-48-H 11.70600-3-48-H 11.16
Specimen
10.4
10.1
9.4
0.1
0.3
9.2
11.3
12.0
11.9
11.3
0.2
0.2
0.2
0.2
0.2
0.3
4.2 85BFailure modes and post-peak ductility
4.2.1 158BShort columns
The loading progression for the 362162S‐33 short columns is depicted in 1427HFigure 5.36
(without a hole) and 1428HFigure 5.37 (with a hole). Both columns exhibit local buckling of the
web near the supports combined with one distortional half‐wave along the length. This
distortional buckling pattern is consistent with that predicted by the elastic buckling
mode shapes of 1429HFigure 5.32a. For the column with the hole, localized hole deformation
( 1430HFigure 5.37, rightmost picture) initiates at a load of approximately 0.4Ptest and increases
in magnitude as the test progresses. This observed deformation behavior is visually
consistent with the “unstiffened strip” approach discussed in Error! Reference source
207
not found., where the strip of web on either side of the hole is assumed to behave as an
unstiffened element.
The inward flange deformation concentrates at the hole after peak load in the short
362S162‐33 specimens with holes. It is hypothesized that the slotted hole reduces the
post‐peak resistance of the web, causing the flanges and lips to carry more of the column
load. This reduction in post peak resistance is quantified by observing the reduction in
area under the load‐displacement curve for the column with the slotted hole, as shown
in 1431HFigure 5.38.
Distortional buckling in one half-wave at peak load
(a) P=0 kips (b) P=7.0 kips (c) P=10.5 kips (peak load)
(d) P=7.5 kips
Figure 5.36 Load‐displacement progression for short column specimen 362‐2‐24‐NH
208
Local buckling at hole (unstiffened strip)
(a) P=0 kips (b) P=10.4 kips(peak load)
(c) P=7.0 kips
Figure 5.37 Load‐displacement progression for short column specimen 362‐2‐24‐H
0 0.05 0.1 0.15 0.20
2
4
6
8
10
12
14
Column axial displacement (inches)
Col
umn
axia
l loa
d (k
ips)
362-2-24-NH362-2-24-H
Slotted hole has small influence on peak load
Slotted hole influences post-peak load path and reduces column ductility
Figure 5.38 Load‐displacement curve for a 362S162‐33 short column with, without a slotted hole
209
Position transducers placed at the mid‐height of the short column specimens capture
the rate of lateral flange displacement associated with distortional buckling, δD, as shown
in 1432HFigure 5.39. 1433HFigure 5.39 demonstrates that the initiation of web local buckling does
not influence the axial stiffness of specimen 362‐2‐24‐NH, but rather that a softening of
the load‐axial deformation curve coincides with the increased rate of lateral flange
movement (distortional buckling). This observation suggests that the loss in axial
stiffness associated with distortional buckling plays a larger role than web local buckling
in the peak load response of the 362S162‐33 short columns. The influence of the slotted
hole on lateral flange displacement is provided in 1434HFigure 5.40, where the post‐peak
flange displacement rates are significantly higher for the 362S162‐33 short column with
holes. The results of 1435HFigure 5.40 indicate that holes potentially have a significant impact
on the collapse mechanisms triggered from distortional buckling. Lateral flange
displacement plots are provided for all specimens in 1436HAppendix F.
210
0 0.05 0.1 0.15 0.20
2
4
6
8
10
12
Col
umn
axia
l loa
d (k
ips)
0 0.05 0.1 0.15 0.20
0.5
1
1.5
2
Column axial displacement (inches)
δ D (inc
hes)
Local buckling half-waves first observed
Rate of flange distortion increases as load-displacement curve softens
westδ+2
eastwestD
δδδ +=
eastδ+
Figure 5.39 Comparison of load‐deformation response and lateral flange displacements for specimen 362‐2‐24‐NH
0 0.05 0.1 0.15 0.20
0.2
0.4
0.6
0.8
1
1.2
1.4
1.6
1.8
2
Column axial displacement (inches)
δ D (inc
hes)
362-2-24-NH362-2-24-H
Increased rate of flange distortion is observed for column with a hole after peak load is reached
Peak load occurs here
NH
H
westδ+2
eastwestD
δδδ +=
eastδ+
Figure 5.40 Influence of a slotted hole on 362S162‐33 short column lateral flange displacement
211
1437HFigure 5.41 and 1438HFigure 5.42 depict the deformation response of the 600S162‐33 short
columns with and without a slotted hole. In both cases, local buckling at the loaded
ends combines with one distortional half‐wave along the column length. The
distortional buckling pattern for these specimens is not wholly consistent with the elastic
buckling predictions of 1439HFigure 5.33a, which shows two distortional half‐waves; however,
specimens 600‐2‐24‐H and 600‐3‐24‐H did buckle in two half‐waves, see 1440HAppendix F for
pictures. These results suggest that geometric imperfections also have a role to play in
the details of the buckling mode initiated in the loaded response. The deformation
response of the member with and without the hole is similar through the test
progression, suggesting that the hole has a small influence on compressive strength and
post‐peak ductility for the hole width to web width ratios considered here. 1441HFigure 5.43
confirms that the slotted hole has a minimal effect on the post‐peak load response of the
column.
212
Web local buckling
Distortional buckling in one half-wavelength at peak load
(a) P=0 kips (b) P=8.0 kips (c) P=11.9 kips (peak load)
(d) P=8.0 kips
Figure 5.41 Load‐displacement progression for short column specimen 600‐1‐24‐NH
Web local buckling initiates
Similar failure mode to no hole specimens
Local buckling near supports combines with distortional buckling (one half-wave) at peak load
(a) P=0 kips (b) P=7.5 kips (c) P=12.1 kips(peak load)
(d) P=8.0 kips
Figure 5.42 Load‐displacement progression for short column specimen 600‐1‐24‐H
213
0 0.05 0.1 0.15 0.20
2
4
6
8
10
12
14
Column axial displacement (inches)
Col
umn
axia
l loa
d (k
ips)
600-1-24-NH600-1-24-H
Slotted hole has small influence on post-peak response and ductility
NH
H
Figure 5.43 Comparison of load‐displacement response for short 600S162‐33 column specimens with and without holes
4.2.2 159BIntermediate length columns
1442HFigure 5.44 and 1443HFigure 5.45 summarize the deformation response of the 362S162‐33
intermediate length columns with and without holes. In both cases, local web buckling
is first observed at approximately 0.45Ptest which is lower than, but the same order of
magnitude as, the calculated local critical elastic buckling load Pcrl. Note in 1444HFigure 5.45
that the local buckling half‐waves are dampened in the vicinity of the holes, similar to
the elastic buckling prediction of 1445HFigure 5.32b. Further, the observed local buckling
waves are at half‐wavelengths consistent with the local buckling mode in 1446HFigure 5.32b,
not those shown interacting with distortional buckling. (This observation supports the
idea that the fundamental elastic buckling modes L, D, and G are representative of the
physical behavior of the column and that the mixed modes observed in an eigenbuckling
analysis only exist numerically.) Three distortional buckling half‐waves become well‐
214
formed at approximately 0.70Ptest, overcoming the local half‐waves in the web except at
the mid‐height of the column. This distortional buckling pattern is consistent with the
elastic buckling prediction in 1447HFigure 5.32b. 1448HFigure 5.46 demonstrates that the presence of
slotted holes has only a minimal influence on load‐axial displacement response.
All of the 362S162‐33 intermediate length columns failed soon after the peak load
with a sudden loss in load‐carrying capacity caused by global flexural‐torsional
buckling. Yielding of the column flanges reduces the torsional stiffness of the section,
and the friction end conditions could not restrain the twisting of the column. The
twisting of specimen 362‐3‐48‐NH is quantified in 1449HFigure 5.47 as the difference between
the west and east mid‐height flange displacements, δT, captured by the position
transducers. The lateral displacement of the flange tips due to distortional buckling (δD),
also shown in 1450HFigure 5.47, is separated from the twisting effect by averaging the west
and east mid‐height flange displacements. 1451HFigure 5.47 shows that the cross‐section is
both ‘opening’ and ‘twisting’, but it is the abrupt increase in δT occurring well past peak
load that leads to the collapse of the member in the test.
215
Web local buckling with flange distortion in three half-waves
Pure distortional buckling dominates over local buckling in this half-wave
Local buckling mixes with distortional half-wave at peak load
1
2
3
(a) P=0 kips (b) P=8.0 kips (c) P=9.5 kips(peak load)
Figure 5.44 Load‐displacement progression, intermediate length column specimen 362‐3‐48‐NH
Hole dampens web local buckling by creating two stiff web strips on either side of hole that locally boost Pcrl above Pcrd
(a) P=0 kips (b) P=8.0 kips (c) P=9.4 kips(peak load)
Figure 5.45 Load‐displacement progression for intermediate length column specimen 362‐3‐48‐H
216
0 0.05 0.1 0.15 0.20
2
4
6
8
10
12
14
Column axial displacement (inches)
Col
umn
axia
l loa
d (k
ips)
362-3-48-NH362-3-48-H
Columns fail abruptly with a global torsionalbuckling mode
NH
H
Figure 5.46 Load‐displacement curve, 362S162‐33 intermediate column with and without a hole
0 0.05 0.1 0.15 0.2-2
-1.5
-1
-0.5
0
0.5
1
1.5
2
Column axial displacement (inches)
δ D, δT (
inch
es)
Peak load occurs here
Abrupt failure
2westeast
D
δδδ +=
2westeast
T
δδδ −=
westδ+ eastδ+
Figure 5.47 362S162‐33 long column mid‐height flange displacements show the global torsional failure mode
217
The load‐displacement response for the intermediate length 600S162‐33 columns
with and without slotted holes is depicted in 1452HFigure 5.48 and 1453HFigure 5.49. Local
buckling is observed at approximately 0.45Ptest for both sections. The holes do not restrict
the local buckling half‐waves as was the case in the 362S162‐33 intermediate length
columns. This local buckling behavior is consistent with that observed in the elastic
buckling analysis, see 1454HFigure 5.33b. Three distortional half‐waves form as the columns
(all 3 of the 600S162‐33 intermediate length specimens) approach peak load. Two loud
sounds resonate from the columns near peak load as the local web buckling half‐waves
at the two column ends abruptly snap into one distortional half‐wave per end. The
change from local‐dominated to distortional‐dominated web buckling is reflected as two
drops in the load‐displacement response near peak load for the 600S162‐33 column
without holes, as shown in 1455HFigure 5.50. The 600S162‐33 column with slotted holes is not
affected by this abrupt mode switching, as it maintains web local buckling well beyond
peak load. The observations suggest that in this case the holes are beneficial because
they maintain the local buckling half‐waves through peak load, allowing the column to
rely more on the post‐peak strength provided by the buckled web. This mode switching
is a difficult challenge for numerical models and these results, repeated in 3 tests,
provides an important and challenging experimental benchmark for the numerical
modeling of these members Section 1456H7.2.
218
Multiple local half-waves change to one distortional half-wave with loud resonant sound at peak load
(a) P=6.0 kips (b) P=11.2 kips(peak load)
(c) P=11.2 kips (d) P=7.0 kips
Figure 5.48 Load‐displacement progression, intermediate length column specimen 600‐1‐48‐NH
Hole preserves web local buckling through peak load
(a) P=0 kips (b) P=11.2 kips(peak load)
(c) P=10.0 kips
Figure 5.49 Load‐displacement progression, intermediate length column specimen 600‐1‐48‐NH
219
0 0.05 0.1 0.15 0.20
2
4
6
8
10
12
14
Column axial displacement (inches)
Col
umn
axia
l loa
d (k
ips)
600-1-48-NH600-1-48-H
Drops in load occur when multiple web local buckling half-waves change abruptly to one distortional half-wave (north end first, then south end)
Figure 5.50 Load‐displacement comparison of intermediate length 600S162‐33 specimens with and without holes
4.3 86BDiscussion of hole influence on elastic buckling and tested response Both local and distortional elastic buckling were observed in the tested response of
the specimens and contributed in different ways to the failure modes of the columns.
Local buckling initiated plastic folding in the web at peak load, and distortional
buckling was reflected as either opening (‐δD) or closing (+δD) of the cross‐section and
yielding of the flanges and lip stiffeners. All three of the short 362S162‐33 columns with
holes exhibited a ‘closed’ distortional buckling failure (+δD), where the presence of the
slotted hole concentrated the plastic deformation in the flanges and lips adjacent to the
hole. This result was different from the short 362S162‐33 columns without holes where
mixed local‐distortional failures were observed. The slotted holes also changed the
220
buckling influence at peak load in the intermediate length 600S162‐33 specimens, where
the holes prevented local web buckling from switching to distortional buckling in all
three specimen tests. The deformation at peak load for the intermediate length 362S162‐
33 and short 600S162‐33 specimens was less sensitive to the presence of slotted holes,
exhibiting mixed local‐distortional failure modes consistent with DSM predictions (L
and D of similar magnitudes) as discussed in Section 1457H5.3.2.3.
The visual observations in this study highlight the complex relationship between
elastic buckling and column failure and the sensitivity of their interaction to the choice
of cross‐section and column length. In the cases of the short 362S162‐33 and
intermediate length 600S162‐33 specimens, it is clearly demonstrated that holes can
influence column deformation and ductility by changing how elastic buckling modes,
local and distortional in this case, affect the axial stiffness and plastic deformation of the
column under load. This data is important in the context of the Direct Strength Method,
especially for this current effort to extend DSM to members with holes, since elastic
buckling is used to predict the failure mode (local, distortional, or global) and ultimate
strength .
4.4 87BDiscussion of friction-bearing boundary conditions
The friction‐bearing end conditions used in this testing are advantageous because
specimen alignment and preparation can be performed without welding or the use of
grout or hydrostone. The specimens were aligned by hand in the testing machine
without special equipment. However, preparing the specimen ends with a milling
221
machine can be time consuming. Further, small deviations in flatness may significantly
impact the tested results and failure modes; real care must be taken in the specimen end
preparation. Finally, lack of a positive connection between specimen and platen makes
it difficult to exactly know the boundary conditions.
In this study, friction between the column ends and the platens prevented a change
in shape of the cross‐section up to peak load in all specimens, but slipping of the cross‐
section was observed after peak load. This slipping was signaled by loud metal‐on‐
metal “popping” sounds associated with observable changes in the cross‐section ( ‐δD of
the flanges, see 1458HFigure 5.47 for definition) at the column ends. Also, uplift warping
deformations like those shown in 1459HFigure 5.51 occurred in the post‐peak range for the
short 600S162‐33 columns experiencing distortional type failures. Distortional buckling
modes are anticipated to be sensitive to this uplift since they are highly sensitive to
warping deformations. The intermediate length 362S162‐33 columns experienced a
sudden global flexural‐torsional failure shortly after reaching peak load as the twisting
of the columns overcame the friction between the column ends and the platens. The
friction‐bearing end conditions did not allow a detailed study of the global flexural‐
torsional post‐peak response for the intermediate length 362S162‐33 columns and likely
decreased their ultimate strengths.
Overall, for short and intermediate length column testing focused on local and
distortional buckling modes, the advantages of the simple friction‐bearing boundary
conditions outweighed the disadvantages. Proper care must be taken to insure the ends
are milled flat and the platens are level and parallel. For longer column tests, where
222
large torsional rotations must be restrained, the bearing conditions employed here are
not recommended for use.
Flange-lip corner lifts off platen when large deformations exist past peak load
Figure 5.51 Short 600S162‐33 column flange‐lip corner lifts off platen during post‐peak portion of test
223
Chapter 6 5BPredicting residual stresses and plastic strains in cold-formed steel members
Thin cold‐formed steel members begin as thick, molten, hot steel slabs. Each slab is
typically hot‐rolled, cold‐reduced, and annealed before coiling and shipping the thin
steel sheet to roll‐forming producers (US Steel 1985). Once at a plant, the sheet is
unwound through a production line and plastically folded to form the final shape of a
structural member, as shown in 1460HFigure 6.1. This manufacturing process imparts residual
stresses and plastic strains through the sheet thickness. These residual stresses and
strains influence the load‐displacement response and ultimate strength of cold‐formed
steel members.
In previous work a statistical approach was employed to draw conclusions on the
magnitude and distribution of longitudinal residual stresses using a data set of surface
strain measurements collected by researchers between 1975 and 1997 (Schafer and Peköz
1998). The measured surface strains are converted to residual stresses using Hooke’s
224
Law and then distributed through the thickness as membrane (constant) and bending
(linear variation) components. These residual stress distributions are a convenient way
to express the measured residual surface strains, and are convenient as well for use in
nonlinear finite element analyses, but they are not necessarily consistent with the
underlying mechanics.
Figure 6.1 Cold‐formed steel roll‐forming: (left) Sheet coil enters roll‐forming line, (right) steel sheet is cold‐formed into C‐shape cross‐section (photos courtesy of Bradbury Group).
Plastic bending, followed by elastic springback, creates a nonlinear through‐
thickness residual stress distribution, in the direction of bending, as shown in 1461HFigure 6.2
(Shanley 1957). The presence of nonlinear residual stress distributions in cold‐formed
steel members has been confirmed in experiments (Key and Hancock 1993) and in
nonlinear finite element modeling of press‐braking steel sheets (Quach et al. 2006 ). A
closed‐form analytical prediction method for residual stresses and equivalent plastic
strains from coiling, uncoiling, and mechanical flattening of sheet steel has also been
proposed (Quach et al. 2004 ). The same plastic bending that creates these residual
stresses also initiates the cold‐work of forming effect, where plastic strains increase the
apparent yield stress in the steel sheet (and ultimate strength in some cases) (Yu 2000).
225
Together, these residual stresses and plastic strains comprise the initial material state of
a cold‐formed steel member.
Plastic bending
Elastic springback
ε
σ
Elastic springback
Plastic bending
+ =
Plastic bending Elastic springback Nonlinear residual stress distribution
compression
tension
Figure 6.2 Forming a bend: plastic bending and elastic springback of thin sheets results in a nonlinear through‐thickness residual stress distribution.
A general method for predicting the manufacturing residual stresses and plastic
strains in cold‐formed steel members is proposed here. The procedure is founded on
common industry manufacturing practices and basic physical assumptions. The
primary motivation for the development of this method is to define the initial state of a
cold‐formed steel member for use in a subsequent nonlinear finite element analysis. The
derivation of the prediction method is provided for each manufacturing step, and the
predictions are evaluated with measured residual strains from existing experiments.
The end result of the method is intended to be accessible to a wide audience including
manufacturers, design engineers, and the academic community. This method also has
the potential to compliment and improve Chapter A7.1.2 of the existing Specification
(AISI‐S100 2007), which currently allows for an increase in member strength from the
226
cold‐work of forming effect at cross‐section corners, but does not directly account for the
influence of the nonlinear through‐thickness corner residual stresses or the influence of
plastic strains and residual stresses from coiling, uncoiling, and flattening of the sheet
steel.
6.1 38BStress‐strain coordinate system and notation
The stress‐strain coordinate system and geometric notation used in the forthcoming
derivations are defined in 1462HFigure 6.3. The x‐axis is referred to as the transverse direction
and the z‐axis as the longitudinal direction of a structural member. Cross‐section
elements are referred to as either “corners” or “flats”. The sign convention for stress and
strain is positive for tension and negative for compression.
Forming direction A
A
sheet steel coil
roller dies
xy
z
z
xy
rz
t
Section A-A
Elevation View
rx
Figure 6.3 Stress‐strain coordinate system as related to the coiling and cold‐forming processes.
227
6.2 39BPrediction method assumptions
The following assumptions are employed to develop this prediction method:
a. Plane sections remain plane before and after cold‐forming of the sheet steel. This
assumption permits the use of beam mechanics to derive prediction equations.
b. The sheet thickness t remains constant before and after cold‐forming of the sheet
steel. A constant sheet thickness is expected after cold‐bending if the bending is
performed without applied tension (Hill 1950). Cross‐section measurements
demonstrate modest sheet thinning at the corners, where t in the corners is typically five
percent less than in the flange and web (Dat 1980). This thinning is ignored here to
simplify the derivations, although a reduced thickness based on the plastic strain
calculations in Section 1463H6.4 could be used if a higher level of accuracy is required.
c. The sheet neutral axis remains constant before and after cross‐section cold‐forming.
Theoretical models used in metal forming theory do predict a small shift in the through‐
thickness neutral axis towards the inside of the corner as the sheet plastifies (Hill 1950).
This shift is calculated as six percent of the sheet thickness, t, when assuming a
centerline corner radius, rz, of 2.5t. A neutral axis shift of similar magnitude has been
observed in the nonlinear finite element model results for thin press‐braked steel sheets
(Quach et al. 2006 ). This small shift is ignored here to simplify the derivations.
228
d. The steel stress‐strain curve is assumed as elastic‐perfectly plastic when calculating
residual stresses. More detailed stress‐strain models that include hardening are
obviously possible, but a basic model is chosen to simplify the derivations. The
implication of this assumption is that the residual stresses may be underestimated,
especially in corner regions where the sheet has yielded completely through the
thickness.
e. Plane strain behavior is assumed to exist during coiling, uncoiling, and flattening
(εx=0) and during cross‐section cold‐forming (εz=0).
f. The steel sheet is fed from the top of the coil into the roll‐forming bed as shown in
1464HFigure 6.4a. This assumption is consistent with measured bending residual stress data
(see Section 1465H6.6) and manufacturing setups suggested by roll‐forming equipment
suppliers ( 1466HFigure 6.1). The author did observe the alternative setup in 1467HFigure 6.4b (sheet
steel unrolling from the bottom of the coil) at a roll‐forming plant, suggesting that the
direction of uncoiling is a source of variability in measured residual stress data.
229
Sheet has residual CONCAVE curvature coming off the coil
Sheet has residual CONVEX curvature coming off the coil
(a)
(b)
Roll-forming bed
Figure 6.4 Roll‐forming setup with sheet coil fed from the (a) top of the coil and (b) bottom of coil. The orientation of the coil with reference to the roll‐forming bed influences the direction of the coiling residual
stresses.
g. Membrane residual stresses are zero. Membrane residual stresses have been
measured by several researchers (Ingvarsson 1975; Dat 1980; Weng and Peköz 1990; De
Batista and Rodrigues 1992; Kwon 1992; Bernard 1993; Key and Hancock 1993), although
the magnitudes are small relative to bending residual stresses (see 1468HTable 6.1). Membrane
residual stresses are experimentally determined by averaging the measured surface
strains on the two faces of a thin steel sheet. Given the variability inherent in these
measurements it is difficult to know if the resulting membrane stresses (strains) are real
or simply unavoidable experimental error.
230
6.3 40BDerivation of the residual stress prediction method
The prediction method proposed here assumes that two manufacturing processes
contribute to the through‐thickness residual stresses in cold‐formed steel members: (1)
sheet coiling, uncoiling, and flattening, and (2) cross‐section roll‐forming. Algebraic
equations for predicting the through‐thickness residual stress and effective plastic
strains in corners and flats are derived here and then summarized in flowcharts in
1469HFigure 6.13 and 1470HFigure 6.17.
3.1 88BResidual stresses from sheet coiling, uncoiling, and flattening
Coiling the sheet steel after annealing and galvanizing, but prior to shipment, may
yield the steel if the virgin yield strain, εyield, is exceeded. If plastic deformation does
occur, a residual curvature will exist in the sheet as it is uncoiled. This residual
curvature is locked into a structural member resulting in longitudinal residual stresses
as the sheet is flattened by the roll‐formers. This process of coiling, uncoiling with
residual curvature, and flattening is described in 1471HFigure 6.5.
231
Coiling Uncoiled with residual curvature
Flattened as sheet enters the roll-formers
Change in curvature locks in bending residual stresses in final member
DETAIL A
DETAIL A
Figure 6.5 Coiling of the steel sheet may result in residual curvature which results in bending residual stresses as the sheet is flattened.
3.1.1 160BCoiling
The through‐thickness strain induced from coiling is related to the radial location of
the sheet in the coil rx, with the well known relationship from beam mechanics:
x
z
ry1
=ε
. (6.1)
εz is the engineering strain through the thickness y in the coiling (longitudinal) direction
z. y varies from ‐t/2 to t/2, where t is the sheet thickness. The radius associated with the
elastic‐plastic threshold initiating through‐thickness yielding from coiling, rep, is derived
by substituting εz=εyield and y=t/2 (outer fiber strain) into Eq. 1472H(6.1):
yield
ep
trε2
= . (6.2)
When the coil radius rx is greater than rep the sheet steel experiences only elastic
deformation on the coil. For sheet steel rolled to a coil radius rx less than rep, through‐
thickness yielding will occur as shown in 1473HFigure 6.6.
232
+σyield
-σyield
y
zc
Figure 6.6 Longitudinal residual stress distribution from coiling.
When rx < rep, the depth of the elastic core c is defined as:
trc yieldx ≤= ε2 . (6.3)
3.1.2 161BUncoiling
As the yielded sheet is uncoiled in preparation for the roll‐forming line, the sheet
steel springs back elastically resulting in a change in the through‐thickness stress. This
stress distribution is determined by first calculating the plastic coiling moment
( )⎥⎥⎦
⎤
⎢⎢⎣
⎡−⎟
⎠⎞
⎜⎝⎛= 2
2
31
2 yieldxyieldcoilx rtM εσ , (6.4)
and then applying an opposing moment elastically to simulate the removal of the
imposed radial displacement
3
12t
yM coilxuncoil
z
−=σ . (6.5)
3.1.3 162BFlattening
After the sheet has been unrolled, a permanent radius of curvature will still exist if rx
was less than rep on the coil. This permanent radius is
233
EIM
r
r coilx
x
x
uncoil
−=
11
. (6.6)
Steel sheet with permanent curvature from coiling is pressed flat as the sheet enters the
roll‐forming line. The longitudinal stresses resulting from flattening the sheet are
simply
uncoil
x
flattenz r
yE−=σ . (6.7)
3.1.4 163BResidual stress distribution
The total through‐thickness longitudinal residual stress distribution due to coiling,
uncoiling, and flattening is presented in 1474HFigure 6.7.
zσ
+ + =
y
z
Coil Uncoil Flatten
+σyield
-σyield
y
zc
flattenzσ
z
y
uncoilzσ
z
y
Residual Stress
Figure 6.7 Predicted longitudinal residual stress distribution from coiling, uncoiling, and flattening of a steel sheet.
The resulting residual stress, σz, is self‐equilibrating for axial force through the thickness
but causes a residual longitudinal moment. Section 1475H6.6 compares the stresses caused by
this moment with surface strains (stresses) measured in experiments.
The longitudinal residual stresses also will create transverse stresses across the
width of the coil, assuming plane strain conditions for an infinitely wide sheet.
234
Supporting the plane strain assumption is the observation that while the actual width of
the sheet is finite, it remains several orders of magnitude greater than the sheet
thickness. Under this assumption, and further assuming only elastic stresses, the
transverse stresses are:
( )flattenuncoilcoilx zzz
σσσνσ ++= . (6.8)
Poisson’s ratio, ν, is assumed here as 0.30 for steel deformed elastically. The through‐
thickness deformation from the uncoiling and flattening components will occur
elastically, and the coiling component will be at least partially elastic through the
thickness for the range of sheet thicknesses common in industry.
3.2 89BResidual stresses from cross-section roll-forming
A set of algebraic equations is derived here to predict the transverse and
longitudinal residual stresses created by roll‐forming a cross‐section. Roll‐forming
residual stresses are cumulative with the coiling residual stresses derived in Section 1476H6.3.1
and provide a complete prediction of the initial stress state of the member cross‐section.
The roll‐forming residual stresses are assumed to exist only at the location of the formed
corners, between the roller die reactions, as shown in 1477HFigure 6.8. Some yielding is
expected to occur outside of the roller reactions as the stress distribution transitions
from fully plastic to fully elastic; however, this transition area is not considered here to
simplify the derivation.
235
Roller Die (Typ.)
Assume bend is fully plastic between roller dies
Figure 6.8 Cold‐forming of a steel sheet.
The engineering strain in the steel sheet, εx, and the bend radius, rz, are related for
both small and large deformations with the strain‐curvature relationship
yrx
z
ε=
1. (6.9)
This geometric relationship is valid for elastic and plastic bending of the steel sheet. For
the small bend radii common in the cold‐formed steel industry (rz =2t to 8t), the steel
sheet yields through its thickness during the cold‐forming process. The steel sheet will
reach the fully plastic stress state shown in 1478HFigure 6.9 as the corner approaches its final
manufactured radius.
-σyield
+σyield
x
y
Figure 6.9 Fully plastic transverse stress state from cold‐forming.
236
After the sheet becomes fully plastic through its thickness, the engineering strain
continues to increase as the radius decreases. When the final bend radius is reached and
the imposed radial displacement is removed, an elastic springback occurs that elastically
unloads the corner (see 1479HFigure 6.2). The change in stress through the thickness from this
elastic rebound is derived with the plastic moment force couple shown in 1480HFigure 6.10.
+σyield
-σyield
t/2 Fp
Fp
x
y
Figure 6.10 Force couple (Fp∙½t) applied to simulate the elastic springback of the steel sheet after the imposed radial deformation is removed.
The plastic moment is calculated with the equation
422
12
2ttttFM yieldyieldP
bendz
σσ=
⋅⋅== , (6.10)
which is then applied elastically through the thickness to simulate the stress distribution
from elastic rebound of the sheet steel:
=⋅⋅
⎟⎟⎠
⎞⎜⎜⎝
⎛
==3
2
1121
4
t
yt
IyM
yield
bendzrebound
x
σ
σ t
yyieldσ3. (6.11)
The final transverse stress state is the summation of the fully plastic stress
distribution through the thickness and the unloading stress from the elastic springback
of the corner as shown in 1481HFigure 6.11, where σx is the transverse residual stress through
237
the thickness from the cold‐forming of the corner. This stress is nonlinear through the
thickness and is self‐equilibrating, meaning that axial and bending sectional forces are
absent in the x‐direction after forming.
+ =
+1.5σyield
-1.5σyield-0.5σyield
+σyield
-σyield
-σyield
+σyield
+0.5σyield
x
y
Plastic Bending Elastic Springback Transverse Residual Stressreboundxσ
xσbendxσ
Figure 6.11 Cold‐forming of a steel sheet occurs as plastic bending and elastic springback, resulting in a self‐equilibrating transverse residual stress.
The transverse residual stresses will create stress in the longitudinal direction due to the
assumed plane strain conditions (see Section 1482H6.2):
xz νσσ = . (6.12)
The Poisson’s ratio, ν, is assumed as 0.30 for steel deformed elastically and 0.50 for fully
plastic deformation. The longitudinal residual stresses through the thickness, σz, are
determined based on these assumptions as shown in 1483HFigure 6.12. Longitudinal residual
stress, σz, is self‐equilibrating for axial force through the thickness but causes a residual
longitudinal moment. This moment is hypothesized to contribute to the observed
longitudinal residual strains measured in experiments (refer to Section 1484H6.6 for a
comparison of this prediction to actual measurements).
238
-0.05σyield
+0.05σyield
+0.50σyield-0.50σyield+ =
+1.5σyield
-1.5σyield
-σyield
+σyield
0.50 0.30 z
y
Plastic Bending Elastic Springback Longitudinal Residual Stressbendxplasticσν rebound
xelasticσνzσ
Figure 6.12 Plastic bending and elastic springback from cold‐forming in the transverse direction result in longitudinal residual stresses because of the plane strain conditions.
A flowchart summarizing the proposed prediction method for residual stresses in
roll‐formed members is provided in 1485HFigure 6.13. 1486HFigure 6.13 explicitly demonstrates
how coiling, uncoiling, flattening, and roll‐forming contribute to the residual stresses
locked into the cross‐section during manufacturing.
239
Start
Yielding on the Coil?
Flat or Corner?
Yielding on the Coil?End
Flat Corner
Yes
No
Yes
No
End
End
yieldx
trε2
≤
yieldx
trε2
>
Yes, yields on coil
No, remains elastic
yieldx
trε2
≤
yieldx
trε2
>
Yes, yields on coil
No, remains elastic
yieldxrc ε2=
IyM coil
xuncoilz
−=σ
( ) ⎥⎦
⎤⎢⎣
⎡−⎟
⎠⎞
⎜⎝⎛= 2
2
31
2 yieldxyieldcoilx rtM εσ 31
121 tI ⋅⋅=
EIM
r
r coilx
x
x
uncoil
−=
11
uncoilx
flattenz r
yE−=σ
yieldσ+ 22tyc
≤≤
xryE
22cyc
<<−
yieldσ−22cyt
−≤≤−
=coilzσ
yieldσ+ 22tyc
≤≤
xryE
22cyc
<<−
yieldσ−22cyt
−≤≤−
=coilzσ
Sheet Coiling
Sheet Uncoiling
Sheet Flattening
Corner Bending
Corner Rebound
Sheet Coiling
Sheet Uncoiling
Sheet Flattening
bendxplastic
bendz σνσ =
tyyieldrebound
x
σσ
3= 22
tyt≤≤−
No residual stresses!
yieldσ−
02
≤≤− yt=bend
xσ 20 ty ≤≤
yieldσ+
50.0=plasticν
reboundxelastic
reboundz σνσ = 30.0=elasticν
coilzelastic
coilx σνσ = 30.0=elasticν
uncoilzelastic
uncoilx σνσ = 30.0=elasticν
flattenzelastic
flattenx σνσ = 30.0=elasticν
(For rz<8t)
Figure 6.13 Flowchart summarizing the prediction method for residual stresses in roll‐formed members.
240
6.4 41BDerivation of effective plastic strain prediction method
In the method proposed here, plastic strains occur from sheet coiling and cold‐
forming, and together with residual stresses describe the initial material state of the
member. The general state of plastic strain at a point can be quantified by using the von
Mises yield criterion extended to plastic deformations (Chen and Han 1988):
( )23
22
213
2 εεεε ++=p (6.13)
where εp is the effective plastic strain, and ε1, ε2,and ε3 are the principal strains. All of the
strains are “true” strains, which may be calculated from the engineering strains via:
)1ln(1 xεε += , )1ln(2 yεε += , )1ln(3 zεε += , (6.14)
where εx, εy, εz are in the Cartesian coordinate system (1487HFigure 6.3) and x,y,z is coincident
with the principal directions. True strains are employed instead of engineering strains to
accommodate the large deformations from plastic bending. Also, from a practical
standpoint, nonlinear FE codes such as ABAQUS (ABAQUS 2007a) require the engineer
to provide true stress, true strain information (as large deformation theory is employed).
The steel sheet is assumed to remain incompressible while experiencing plastic
deformations, therefore when calculating εp
0321 =++ εεε . (6.15)
241
4.1 90BEffective plastic strain from sheet coiling
Engineering plastic strains, as shown in 1488HFigure 6.14, accumulate during the coiling of
sheet steel if the coiling radius rx is less than the elastic‐plastic threshold rep.
c z
y
pzε
Figure 6.14 Plastic strain distribution from sheet coiling with a radius less than elastic‐plastic threshold rep.
The engineering plastic strain distribution from coiling is:
2
, cyry
yieldx
pz ≥−= εε
2
, cyry
yieldx
pz −≤−= εε (6.16)
otherwisepz 0=ε ,
where the elastic core, c, is defined in Eq. 1489H(6.3). Plane strain conditions result in ε1=0, and
ε2=‐ε3 via the incompressibility assumption of Eq. 1490H(6.15). Further, the Cartesian
coordinate system is coincident with the principal axes, resulting in the following true
principal plastic strains:
01 =ε , )1ln(2pzεε +−= , )1ln(3
pzεε += . (6.17)
Substituting the principal strains into Eq. 1491H(6.13) and simplifying leads to the through‐
thickness effective plastic strain from coiling
242
( )pz
coilingp εε += 1ln
32
. (6.18)
This plastic strain distribution, depicted in 1492HFigure 6.15, will exist at all locations in the
cross‐section (corners and flats) when rx is less than the elastic‐plastic threshold rep.
y
c
coilingpε
Figure 6.15 Effective plastic strain in a cold‐formed steel member from sheet coiling when the radius rx is less than the elastic‐plastic threshold rep.
The plastic strain from coiling, εpcoiling, will generally be much smaller in magnitude than
the plastic strain from cross‐section cold‐forming, εpbend, as discussed in following section.
4.2 91BEffective plastic strain from cross-section cold-forming
Large transverse plastic strains occur through the thickness of a thin steel sheet when
the sheet is permanently bent. The engineering plastic strain distribution from cold‐
forming is described via
z
px r
y−=ε , (6.19)
which assumes that the elastic core at the center of the sheet is infinitesimally small.
This assumption is consistent with the small bend radii common in industry (see 1493H6.3.2).
243
Plane strain conditions and Eq. 1494H(6.15) result in ε3=0, ε2=‐ε1. Physically these conditions
imply that the sheet will experience some thinning at the location of cold‐forming (see
Section 1495H6.2), but the tendency to plastically shorten longitudinally will be resisted by the
adjacent undeformed portion of the cross‐section. As before, the Cartesian coordinate
system is coincident with the principal axes, resulting in the following plastic principal
strains:
)1ln(1pxεε += , )1ln(2
pxεε +−= , 03 =ε . (6.20)
Substituting for the principal strains and simplifying, the effective plastic strain at a
cold‐formed corner is:
( )px
bendp εε += 1ln
32
(6.21)
This effective plastic strain distribution is shown in 1496HFigure 6.16. The distribution exists
only at the cold‐bent locations in a cross‐section and should be added to the coiling
plastic strain distribution in 1497HFigure 6.15.
y
Figure 6.16 Effective von Mises true plastic strain at the location of cold‐forming of a steel sheet.
A flowchart summarizing the prediction method for effective plastic strains in roll‐
formed members is provided in 1498HFigure 6.17.
244
Start
Yielding on the Coil?
Flat or Corner?
Yielding on the Coil?End
Flat Corner
Yes
No
Yes
No
End
End
yieldx
trε2
≤
yieldx
trε2
>
Yes, yields on coil
No, remains elastic
yieldx
trε2
≤
yieldx
trε2
>
Yes, yields on coil
No, remains elastic
Sheet Coiling
Corner Bending
Sheet Coiling
No equivalent plastic strains!
( )pz
coilingp εε += 1ln
32
yieldxrc ε2=
( )px
bendp εε += 1ln
32
22tyt
≤≤−
=pzε 2
cy −≤
2cy ≥yield
xry ε−
0 otherwise
yieldxry ε−
z
px r
y−=ε
Figure 6.17 Flowchart summarizing the prediction method for effective plastic strains in roll‐formed members
6.5 42B Employing the prediction method in practice: quantifying the coil radius influence
The residual stress and plastic strain distributions derived for cross‐section cold‐
forming (Sections 1499H6.3.2 and 1500H6.4.2) are straight‐forward to calculate if the yield stress, σyield,
and thickness, t, of the sheet steel are known. The coiling residual stresses and plastic
strains are more difficult to calculate because the coil radius coinciding with the as‐
formed member, i.e., the radial location of the sheet, rx, is almost always unknown in
245
practice. However, rx can be derived in an average sense though, since the range of inner
and outer coil radii are known and the probability that a structural member will be
manufactured from a certain rx can be quantified.
The relationship between coil radius, rx, and corresponding linear location S of the
sheet within the coil can be described using Archimedes spiral (CRC 2003)
( )22innerx rr
tS −=
π . (6.22)
The spiral maintains a constant pitch with varying radii, where the pitch is the thickness
of the steel sheet, t, as shown in 1501HFigure 6.18, L is the total length of sheet in the coil, and
rinner and router are the inside and outside coil radius, respectively. As‐shipped outer coil
radii range from 24 in. to 36 in. and inner coil radii range from 10 in. to 12 in. These
ranges were determined by the author during a visit to a local roll‐forming plant.
rx
StartEnd
t LS =0=S
S
Figure 6.18 Coil coordinate system and notation.
Archimedes spiral is used to describe the probability that the steel sheet will come
from a certain range of radial locations in the coil. The cumulative distribution function
246
(CDF), FR(rx) = probability that the radius is less than rx, is obtained by normalizing S by
L
22
22
)(innerouter
innerxxR rr
rrrFLS
−−
== . (6.23)
The probability density function (PDF) of rx is calculated by taking the derivative of FR(rx)
22
2)()(innerouter
x
x
xRxR rr
rdr
rdFrf−
== . (6.24)
The mean value of the radial location for a given inner and outer coil radii is
⎟⎟⎠
⎞⎜⎜⎝
⎛+
−+== ∫outerinner
outerinnerouterinner
r
r xxxRx rrrrrrdrrrfr outer
inner 32)( . (6.25)
The variance of the radial location is
( )( )( )2
22222 4
181))((
innerouter
innerouterinnerinnerouterouter
r
r xxxRR rrrrrrrrdrrrrfs outer
inner +−
++=−= ∫ . (6.26)
These statistics for rx can then be used with the prediction method for coiling, uncoiling,
and flattening residual stresses and plastic strains described in Sections 1502H6.3.1 and 1503H6.4.1.
1504HFigure 6.19 summarizes the influence of sheet thickness and virgin yield stress on the
longitudinal residual stress distributions in flats and corners. (The method proposed in
this chapter provides residual stresses and strains for the entire member, only the
longitudinal residual stresses are shown in 1505HFigure 6.19.) The solid lines in 1506HFigure 6.19
are calculated using the mean value, xr =18.7 in, from Eq. 1507H(6.25) assuming rinner=12 in. and
router=24 in. The distributions with the dashed lines are calculated with xr ±sR, where
sR=3.4 in. is calculated with Eq. 1508H(6.26). The residual stresses are nonlinear through the
thickness and have different shapes for flats and corners. The stress magnitudes at the
247
outer fibers increase for thicker sheets and lower yield stresses. The accuracy of the
linear bending residual stress model commonly employed in finite element analyses is
perhaps sufficient when yield stress is low and thickness is high (relatively), but for
typical thicknesses (0.0346 in. to 0.0713 in.) and yield stress (50 ksi) the assumption of a
linear longitudinal stress distribution is not consistent with the mechanics‐based
predictions in 1509HFigure 6.19.
-t/2
t/2flat
-σyield
σyield0
corner flat corner flat corner
-t/2
t/2
-σyield
σyield0
-t/2
t/2
-σyield
σyield0
σyield=30 ksi
t=0.0713 in.
t=0.0346 in.
t=0.1017
σyield=50 ksi σyield=80 ksi
Longitudinal Residual Stresses
Positive stress is tension, negative stress is compression
z
z
z
y
y
y
Figure 6.19 Influence of sheet thickness and yield stress on through‐thickness longitudinal residual stresses (z‐direction, solid lines are predictions for mean coil radius, dashed lines for mean radius +/‐ one standard
deviation).
248
6.6 43B Comparison of prediction method to measured residual stresses
The flat and corner residual surface strain measurements from 18 roll‐formed
specimens are used to evaluate the proposed residual stress prediction method. The
prediction method provides the complete through‐thickness longitudinal strain (stress)
distribution if the radial location in the coil from which the specimen originated in the
coil, rx, is known. Since the radial coil location of the 18 specimens is unknown, rx is
statistically estimated for each specimen using the coil radius that best fits the predicted
surface strains to the measured surface strains from a specimen cross‐section (for both
corners and flats). Once the best fit radial locations have been calculated, they are
examined to determine if their magnitude is rational when compared to typical inner
and outer dimensions of a sheet coil. Although this comparison only provides a partial
evaluation of the prediction method, it is as far as one can go with the available data.
Qualitatively the prediction method is consistent with the more detailed through
thickness findings (Key and Hancock 1993; Quach et al. 2006 ).
6.1 92B Measurement statistics
The mean and standard deviation of the residual stresses for the 18 roll‐formed
specimens used in this comparison are provided in 1510HTable 6.1. Positive membrane
stresses are tensile stresses and positive bending stresses cause tension at y=‐t/2 (see
1511HFigure 6.3 for coordinate system). The statistics demonstrate that both membrane and
bending residual stress measurements are highly variable and that the membrane
249
stresses are small relative to the steel yield stresses. Details on the residual stress
measurements for each of the 18 specimens are described in a previous research
progress report (Moen and Schafer 2007b).
Table 6.1 Statistics of the residual stresses in roll‐formed members
Mean STDEV Mean STDEVCorners 5.7 10.1 32.0 23.8 23Flats 1.8 10.7 25.2 20.7 120
ElementResidual stress as %σyield No. of
SamplesMembrane Bending
6.2 93BMean-squared error (MSE) estimate of radial location
To explore the validity of the prediction method, the flat and corner residual stress
measurements from the 18 specimens are used to estimate the radial location rx from
which each specimen originated. These estimated radial locations are then used to
calculate the difference between the predicted and measured longitudinal residual
stresses.
6.2.1 164BMSE minimization
The location of the specimen in the coil, rx, is estimated by minimizing the sum of the
mean‐squared errors (MSE) for the p=1,2,…nq measurements taken around the cross‐
section of the q=1,2,…,18 specimens
2
1 ,, minargˆ ∑
=⎟⎟⎠
⎞⎜⎜⎝
⎛ −=
qn
p pqyield
predictedpq
measuredpq
qxr σσσ
. (6.27)
Both corner and flat measurements are included in the minimization.
250
6.2.2 165BBending component of longitudinal residual stress distribution
The bending component of the predicted residual stress distribution must be isolated
to compare with the measured values. The total predicted longitudinal residual stress
distribution in the flats and corners of each cross‐section is integrated to calculate the
sectional moment through the thickness
∫−= 2
2
t
t zx ydyM σ . (6.28)
Mx is then converted into a predicted outer fiber bending residual stress which can then
be directly compared to the measurements
I
tM xpredictedpq
⎟⎠⎞
⎜⎝⎛
= 2σ . (6.29)
6.2.3 166B Estimated coil radii using MSE
1512HFigure 6.20 demonstrates the mean‐squared error results for de M. Batista and
Rodrigues Specimen CP1 (De Batista and Rodrigues 1992). The radial location that
minimizes the prediction error is 1.60rinner in this case, and is summarized in 1513HTable 6.2 for
all 18 roll‐formed specimens considered. The estimated radial locations fall within the
range of inner and outer coil radii assumed in the prediction (rinner to 2.40rinner) except for
Dat RFC13 which is slightly outside the range at 2.45rinner. The MSE radial location
cannot be determined in the three Bernard specimens (Bernard 1993) since the bending
residual stresses in the flats are predicted to be zero. These three specimens are cold‐
251
formed steel decking with a thin sheet thickness t ranging from 0.022 in. 0.0400 in. and a
relatively high yield stress σyield ranging from 87 ksi to 94 ksi. In this case, the coiling and
uncoiling of the steel sheet will occur elastically as demonstrated in 1514HFigure 6.19.
Measured bending residual stress magnitudes in the flats of the Bernard specimens are
on average 0.03σyield which is consistent with the prediction method.
1 1.5 2 2.5 3 3.50
1
2
3
4
5
6
7
8
rx/rinner
MS
E
Figure 6.20 The mean‐squared error of the predicted and measured bending residual stresses for de M. Batista and Rodrigues (De Batista and Rodrigues 1992), Specimen CP1 is minimized when rx=1.60rinner.
252
Table 6.2 Radial location in the coil that minimizes the sum of the mean square prediction error for roll‐formed members
Researcher Specimen rx estimatein.
de M. Batista and Rodrigues (1992) CP2 12.0de M. Batista and Rodrigues (1992) CP1 16.0
Weng and Peköz (1990) RFC13 18.0Weng and Peköz (1990) RFC14 11.0Weng and Peköz (1990) R13 14.5Weng and Peköz (1990) R14 13.0Weng and Peköz (1990) P3300 19.5Weng and Peköz (1990) P4100 15.0Weng and Peköz (1990) DC-12 23.0Weng and Peköz (1990) DC-14 16.0
Dat (1980) RFC14 20.0Dat (1980) RFC13 24.5
Bernard (1993) Bondek 1 N/ABernard (1993) Bondek 2 N/ABernard (1993) Condeck HP N/A
Abdel-Rahman and Siva (1997) Type A - Spec 1 16Abdel-Rahman and Siva (1997) Type A - Spec 2 16Abdel-Rahman and Siva (1997) Type B - Spec 1 13
rinner=10 in., router=24 in.N/A coiling residual stresses are predicted as zero
6.3 94BStatistical variations between measurements and predictions
The predicted radial locations in 1515HTable 6.2 are now used to calculate the statistical
variations between the experiments and predictions. The bending residual stresses in
the 18 roll‐formed members are calculated using the MSE‐predicted radial location rx
with the residual stress prediction method summarized in 1516HFigure 6.13. The bending
component of the residual stress prediction is then obtained with Eq. 1517H(6.29). The
difference between the predicted and measured residual bending stresses, epq, for the
p=1,2,…,n measurements taken around the cross‐section of the q=1,2,…,18 specimens is
calculated as
pqyield
predictedpq
measuredpq
pqe,σσσ −
= . (6.30)
253
The error histogram for the flat cross‐sectional elements in 1518HFigure 6.21a demonstrates
that the mean difference μe is near zero with a standard deviation se=0.15σyield. The
scattergram in 1519HFigure 6.21b demonstrates the strength of the correlation between the
measurements and predictions in the flats; the solid regression line passes nearly
through zero (y‐intercept=0.05σyield) and has nearly a unit slope (m=0.92). Also, the
majority of the data lies within ± one standard deviation of the estimate, denoted as the
dashed lines in the figure. It is concluded that the prediction method is consistent with
the measured data in the flats.
The corner element error histogram in 1520HFigure 6.22a shows a negative bias of μe=‐
0.16σyield meaning that the predicted residual stresses are generally higher than the
measured values. The standard deviation of the error is large (se=0.19σyield) but is less
than the standard deviation of the corner residual stress measurements in 1521HTable 6.1
(sm=0.24σyield). This demonstrates a greater match between the measurements and
predictions, although more corner residual stress measurements are needed to improve
the strength of this comparison. The scattergram in 1522HFigure 6.22b highlights the
variability in the measured corner data, especially in the region corresponding to
σpredicted=0.4σyield, where bending residual stresses (strains) vary from 0 to 0.7σyield.
254
(a) (b)
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8-0.2
-0.1
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
σm
easu
red / σ
yiel
d
σpredicted/σyield
-1 -0.5 0 0.5 10
5
10
15
20
25
30
35
40
(σmeasured-σpredicted)/σyield
Obs
erva
tions
yielde
yielde
σσ
σμ
15.003.0
=
=
Figure 6.21 (a) Histogram and (b) scattergram of bending residual stress prediction error (flat cross‐sectional elements) for 18 roll‐formed specimens.
(a) (b)
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8-0.2
-0.1
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8σ
mea
sure
d / σyi
eld
σpredicted/σyield
-1 -0.5 0 0.5 10
5
10
15
20
25
30
35
40
(σmeasured-σpredicted)/σyield
Obs
erva
tions yielde
yielde
σσ
σμ
19.016.0
=
−=
Figure 6.22 (a) Histogram and (b) scattergram of bending residual stress prediction error (corner cross‐sectional elements) for 18 roll‐formed specimens.
6.7 44BDiscussion
The residual stresses and strains predicted with this method (Section 1523H6.3 for stress,
Section 1524H6.4 for strain) form the initial material state in the cross‐section. In design, this
initial material state is sometimes considered through the so‐called cold‐work of forming
255
effect, where the yield stress of the material is increased above the virgin yield stress,
σyield, to account for the ‘working of the corners’. For one‐dimensional stress‐strain this
concept is expressed as shown in 1525HFigure 6.23, where ‘working the corners’ results in a
residual plastic strain, εp, such that when the section is re‐loaded the stress at which
yielding re‐initiates, σey, is greater than the virgin yield stress, σyield. If no residual stresses
existed the apparent increase in the yield stress from σyield to σey can be significant.
However, as 1526HFigure 6.19 illustrates, ‘working the corners’ also contributes to residual
stresses, σers, and these residual stresses may decrease the apparent yield stress.
The prediction method presented herein provides a more nuanced understanding of
the cold‐work of forming effects. The residual plastic strains may increase the apparent
yield stress, but those strains vary through the thickness and have contributions from
both transverse and longitudinal strains. Further, residual stresses follow their own
relatively complicated distribution through the thickness. In a multi‐axial stress state
using the von Mises yield criterion, 1527HFigure 6.23 is enforced for the effective stress –
effective strain pair for every point in the cross‐section. As a result, the apparent yield
stress upon loading varies through the thickness and is influenced by both the residual
stresses and strains. Even under simple loading conditions (e.g., compression) a cold‐
formed member undergoes plate bending well in advance of collapse, so the strains
demanded of the material also vary through the thickness and around the cross‐section.
While it is indeed possible to model such effects in a finite element analysis, assuming
these effects can be collapsed into a generic increase in the yield stress for the entire
section as is currently done in design would seem to be an oversimplification.
256
pyield εε +
σ
ε
yieldσ
Cold-formed steel
Virgin steel
Apparent yield stress
rseσ
Effective residual stress
eyσ
Figure 6.23 Definition of apparent yield stress, effective residual stress, and effective plastic strain as related to a uniaxial tensile coupon test.
Implementation of the residual stresses and initial plastic strains into a commercial
finite element program such as ABAQUS, where the member is modeled using shell
elements, is relatively straightforward. The number of through‐thickness section
(integration) points must be increased to resolve the nonlinear through‐thickness
residual stress and strain distributions. The residual stresses and strains predicted herein
can be relatively large. Further, conventional loading (e.g., compression, major‐axis
bending) may cause loading or unloading of these initial stresses at a given point in the
cross‐section. As a result, the hardening rule: isotropic, kinematic, or mixed can have
practical differences in the observed response even when the applied loads themselves
are not reversing.
For this situation, kinematic hardening, which approximates the Bauschinger effect,
provides a more conservative model of the anticipated material behavior than isotropic
hardening. However, to model kinematic hardening the location of the center of the
yield surface in stress space (also known as the backstress) must be determined for each
257
point in the cross‐section at the end of the manufacturing process. This location is a
function of the extent of yielding, in the example of 1528HFigure 6.23, the backstress would be
the Δσ1, Δσ2, Δσ3 triad that results in the effective stress increasing from σyield to σey.
Unfortunately, the elastic‐perfectly plastic assumption used to predict residual stresses
herein does not directly allow for the calculation of the backstress. However, the
effective plastic strain may be used to approximate the backstress as provided in
1529HAppendix G. Further examination of the predicted residual stress and strains and their
impact on the peak strength and collapse response of cold‐formed steel members in
nonlinear finite element analysis is currently underway, including the work presented in
Section 1530H7.2.
6.8 45BAcknowledgements
The development of this residual stress prediction method would not have been
possible without accurate information about the manufacturing process of sheet steel
coils and cold‐formed steel members. Thanks to Clark Western Building Systems, Mittal
Steel USA, and the Cold‐Formed Steel Engineers Institute (CFSEI) for their important
contributions to this research, especially Bill Craig, Ken Curtis, Tom Lemler, Joe
Wellinghoff, Ezio Defrancesco, Jean Fraser, Narayan Pottore, and Don Allen.
258
Chapter 7 6BNonlinear finite element modeling of cold-formed steel structural members
Commercial finite element programs provide a means for realistic collapse
simulation of cold‐formed steel structural members. Thin shell finite element
formulations provided in ABAQUS (e.g., the S9R5 element discussed in 1531HChapter 2) are
designed to capture the sharp folds and through‐thickness yielding characteristic of
cold‐formed steel beams and columns at their ultimate limit state. Robust solution
algorithms are available to predict unstable, geometrically nonlinear collapse. The
ability to define the initial state of a member, including geometric imperfections and the
effects of residual stresses and initial plastic strains from the manufacturing process, is
also feasible. Care must be taken though with computational results since they are often
sensitive to modeling inputs and assumptions. It is prudent to study these sensitivities
and validate a specific modeling protocol with known experiment results before trusting
the protocol to consistently produce physically realistic results.
259
This chapter begins with preliminary nonlinear finite element studies of stiffened
elements (i.e., a simply supported plate, see 1532HFigure 3.1 for definition) with and without
holes, which are designed to gain experience with available ABAQUS nonlinear finite
element solution methods. The influence of imperfections on stiffened elements is also
evaluated, and the through‐thickness yielding patterns of a stiffened element (i.e.,
“effective width”) with and without a hole are compared. The conclusions reached from
this preliminary work are used to guide the development and validation of a nonlinear
finite element modeling protocol which is needed in 1533HChapter 8 to explore the Direct
Strength Method for members with holes.
7.1 46BPreliminary nonlinear FE studies
Exploratory nonlinear finite element studies are conducted in this section to gain
experience with ABAQUS input parameters and solution controls. All studies are
focused on the simulation of a stiffened element loaded unixaxially to collapse, and
specific attention is paid to the modeling of a stiffened element with a hole. Experience
gained from solving this highly nonlinear problem will be valuable when implementing
the larger simulation studies on full cold‐formed steel members with holes in Section
1534H7.2.
1.1 95BFinite element modeling definitions
The stiffened element is modeled with ABAQUS S9R5 thin shell finite elements,
where the plate dimensions are h=3.4 in. and L=27.2 in. (see 1535HFigure 3.2 for plate
dimension definitions) and the plate thickness t is 0.0346 in. (These dimensions are
260
specifically chosen to be consistent with the flat web width and thickness of an SSMA
362S162‐33 structural stud.) Cold‐formed steel material properties are assumed as
E=29500 ksi and ν=0.30. Material nonlinearity is simulated in ABAQUS with classical
metal plasticity theory, including the assumption of a von Mises yield surface and
isotropic hardening behavior. The nonlinear plastic portion of the true stress‐strain
curve shown in 1536HFigure 7.1 was obtained from a tensile coupon test (Yu 2005) and input
into ABAQUS to define the limits of the von Mises yield surface as a function of plastic
strain.
0 0.02 0.04 0.06 0.08 0.1 0.12 0.14 0.16 0.18 0.20
10
20
30
40
50
60
70
true
plas
tic s
tress
, ksi
true strain
True stress True strainksi
33 0.000034.31 0.001642.14 0.025951.35 0.070060.86 0.1407
65.5 0.1988
Figure 7.1 True stress‐strain curve derived from a tensile coupon test (Yu 2005)
The boundary conditions of the stiffened element are summarized in 1537HFigure 7.2.
The plate is simply‐supported around the perimeter with sides free to wave. The nodes
at the loaded edges of the plate are coupled to displace together longitudinally (in the 1
261
direction), which prevents local failure modes of the plate at the loaded edges. The
nodal coupling is provided with an equation constraint in ABAQUS.
1
2
3
Perimeter restrained in 2 (v=0), unloaded edges free to wave
Transverse midline restrained in 3 (w=0)
Longitudinal midline restrained in 1 (u=0)
Loaded edge coupled to move together in 1 using equation constraint (all u are equal)
Loaded edge coupled to move together in 1 using equation constraint (all u are equal)
Figure 7.2 Simply supported boundary conditions with equation constraint coupling at loaded edges
Two types of loading conditions, uniform load and uniform displacement, are
considered as shown in 1538HFigure 7.3. The uniform compressive load is applied as
consistent nodal loads on the plate edge. The magnitude of the uniform load is
represented by the parameter λ, which is an accumulation of load steps ∆λ automatically
determined by ABAQUS. ∆λ is large when the Newton‐Raphson algorithm converges
quickly (along the linear branch of the load‐displacement curve) and adjusts to smaller
increments as equilibrium becomes more difficult to achieve (near the peak of the load‐
displacement curve). For the uniform displacement case, the total displacement δ of the
plate edges is applied over 100 steps, where the maximum displacement increment at
each step is set to ∆δ=0.0145t.
262
λ Load scaling factor P Unit forceh Loaded width of the plate
Displacement of plate edge
(b)
hPλ
hPλ
δ
(a)
δ
δ
Figure 7.3 Application of (a) uniform load and (b) uniform displacement to a stiffened element
Initial geometric imperfections are imposed based on the fundamental elastic
buckling mode of the stiffened element (see buckled shape in 1539HFigure 7.3). The
magnitude of the imperfections is chosen based on a probabilistic treatment developed
for cold‐formed steel members (Schafer and Peköz 1998). Since the stiffened element
considered here is chosen to be consistent with the web of a structural stud, a Type 1
(local buckling) imperfection is assumed as shown in 1540HFigure 7.4. The maximum
magnitude of the imperfection field is selected such that there is a 50 percent chance that
a randomly occurring imperfection in the plate, ∆, will have a magnitude less than d1,
i.e., P(∆<d1)=0.50. For this probability of occurrence, the initial imperfection field of the
stiffened element is scaled to d1/t=0.34.
263
d1
Cross section with initial geometric imperfections
Figure 7.4 Type 1 imperfection (Schafer and Peköz 1998)
1.2 96BABAQUS nonlinear solution methods
Two nonlinear solution methods, the modified Riks method and a Newton‐Raphson
technique with artificial damping, are available in ABAQUS to solve difficult nonlinear
problems. The modified Riks Method (i.e., *STATIC, RIKS in ABAQUS), was developed
in the early 1980’s and enforces an arc length constraint on the Newton‐Raphson
incremental solution to assist in the identification of the equilibrium path at highly
nonlinear points along the load‐deflection curve. This method is discussed extensively
in several publications (Crisfield 1981; Powell and Simons 1981; Ramm 1981; Schafer
1997; ABAQUS 2007a). Another solution option is a Newton‐Raphson technique (i.e.,
*STATIC, STABILIZE in ABAQUS) which adds artificial mass proportional damping as
local instabilities develop (that is, when changes in nodal displacements increase rapidly
over a solution increment) to maintain equilibrium (Yu 2005; ABAQUS 2007a). Local
instabilities near peak load are common in cold‐formed steel members, such as when a
thin plate develops at a fold line prior to collapse.
264
In this study, the stiffened element described in Section 1541H7.1.1 is loaded to collapse
in ABAQUS employing the modified Riks method with uniform loads applied
uniaxially (see 1542HFigure 7.3a) and then with the artificial damping solution method
employing uniform displacements (see 1543HFigure 7.3b). (Either method is capable of
solving problems with applied loads or applied displacements.) The goal of this
preliminary study is to gain experience with the solution controls for each method.
Additional background information pertaining to the ABAQUS implementation of the
artificial damping method is also discussed to provide specific guidance (and raise
future research questions) on its proper use.
1.2.1 167BModified Riks solution
The load‐displacement curves and deformed shapes (at peak load) of the
stiffened element solved with the modified Riks method are provided in 1544HFigure 7.5.
Different post‐peak equilibrium paths were obtained by varying ∆λmax, the maximum
load increment limit for the ABAQUS automatic step selection algorithm. The existence
of multiple solutions is consistent with a plate containing periodic geometric
imperfections, since each half‐wavelength has an equal chance of deforming locally into
a plastic failure zone.
Although there were several different post‐peak paths observed depending upon
the choice of ∆λmax, the primary failure mechanism for the plate was a sharp yield‐line
fold occurring transversely across the plate. 1545HFigure 7.6 demonstrates that this folding
occurs at the crest of the buckled half‐wave of the initial geometric imperfection field; in
265
this case the failure mechanism of the plate is linked to the initial imperfection shape.
The quantity and location of the plastic folds influenced the overall ductility of the
stiffened element (i.e., the area under the load‐displacement curve). As the number of
folds increase, the post‐peak strength and ductility of the plate increase. The peak
compressive load of the stiffened element was not sensitive to changes in ∆λmax.
0 0.2 0.4 0.6 0.8 10
0.2
0.4
0.6
0.8
1
0 0.25 0.5 0.75 1 1.25 1.5 1.750
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1
δ/t
λP/P
y
max 0.50λΔ =
max 0.10,0.20,0.25,0.30,0.40,0.60
λΔ =
max 0.70λΔ =
max 0.35λΔ =
max 0.05,0.15λΔ =
Failure modes and associated Riks step sizes
Initial imperfection shape (scale exaggerated)
δ
δ
hPλ
hPλ
Figure 7.5 Modified Riks method load‐displacement solutions and failure modes
fold line (typ.)
initial state with assumed
imperfections
ultimate limit state
(imperfection and deformation magnitudes not to scale)
Figure 7.6 Correlation between initial imperfection shape and fold line locations at failure
266
1.2.2 168BArtificial damping solution
7.1.2.2.1 208BBackground on artificial damping solution method
Artificial mass proportional damping is employed in ABAQUS to alleviate local
instabilities in the *STATIC, STABILIZE solution method. The global equilibrium
equations at each displacement step can be written as:
0=−− DFP (7.1)
where P is the vector of applied external forces, F is the vector of calculated internal
forces, and D is the vector of viscous damping forces. The damping force vector D is
calculated at each step based on the following relationship:
vMD )(c= (7.2)
where c is the damping ratio, M is an artificial mass matrix calculated with a unit
material density, and v represents the change in nodal displacements divided by the size
of the “time” step selected by ABAQUS. v is called the “nodal velocity” in ABAQUS
since the dimensions are length/”time”, which makes v sensitive to the definition of
“time”. In this study, the total “time” is selected as one unit and the maximum “time”
step allowed is 0.01 units. If the total “time” is chosen as 100 units and the maximum
“time” step as 1 unit, it seems that the magnitude of the damping forces D would
change. Following the same argument, the magnitude of v is impacted by the choice of
units for the problem (feet, inches, meters, mm) since v has dimensions of length/”time”
units. Future work is needed to evaluate the influence of “time” and length units on the
calculation of the “nodal velocity” v. The evaluation of the artificial damping solution
267
sensitivity to the magnitude and distribution of mass in a member is also another future
research topic.
When the solution is stable, changes in nodal displacements are small and
viscous damping is negligible. When large changes in displacements occur between two
consecutive load steps (as in the case of a local instability), damping forces are applied to
help make up the difference between P and F. v may only be high for certain locations
in the member, and therefore damping will only be applied there. ABAQUS provides
both automatic and manual options for selection the damping factor c; if c is chosen
manually, ABAQUS recommends that it should be chosen as a small number since large
damping forces can add too much artificial stiffness to the system, producing an
unreasonable solution. When the automatic option is selected, ABAQUS finds c such
that the dissipated energy to total strain energy ratio after the first load step is equal to
2.0x10‐4.
1.2.3 169BArtificial damping results
The artificial damping solutions for the stiffened element are presented in 1546HFigure
7.7. Load‐deformation results pertaining to both manually and ABAQUS‐selected
damping factors are plotted, demonstrating that the magnitude of the damping
parameter c influences the post‐peak response and causes the prediction of several
different load paths (in a similar way to how ∆λmax affected the modified Riks solutions).
Peak load is not sensitive to the quantity of damping in this case, and is consistent with
the magnitude predicted with the modified Riks method.
268
0 0.2 0.4 0.6 0.8 10
0.2
0.4
0.6
0.8
1
0 0.25 0.5 0.75 1 1.25 1.5 1.750
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1
δ/t
P/P
y
Initial imperfection shape (scale exaggerated)
δ
δ
Failure modes and associated damping factor c
c=0.000, 0.0162
c=0.005
c=0.050
ABAQUS chooses this path (c=0.0162)
Displacement Control h
P
hP
Figure 7.7 Artificial damping load‐displacement solutions and failure modes
1.3 97BABAQUS nonlinear solution controls
Section 1547H7.1.2 summarizes the preliminary experiences gained using ABAQUS
nonlinear solution methods to determine the ultimate strength of stiffened elements.
Equilibrium paths and failure modes can be sensitive to solution controls, although the
peak resisting load of the plate was consistently predicted. The nonlinear solution of a
stiffened element with a slotted hole is attempted with the modified Riks method
(*STATIC, RIKS), the default Newton‐Raphson solution algorithm (*STATIC), and
Newton‐Raphson with artificial damping (*STATIC, STABILIZE) solution methods
available in ABAQUS. The goal of the study is to determine a set of solutions controls
269
(load step size, damping factor, convergence limits) that is capable of capturing the post‐
peak load‐displacement response of a stiffened element with a hole.
1.3.1 170BProblem description
The stiffened element described in Section 1548H7.1.1 is considered in this study. A
single slotted hole is placed at the midlength of the plate and centered between the
unloaded edges. The slotted hole has dimensions of hhole=1.5 in., Lhole=4 in., and rhole=0.75 in.
(see 1549HFigure 3.2 for hole dimension definitions).
The boundary conditions of the stiffened element were initially assumed to be
simply‐supported with both transverse and longitudinal plate midlines restrained and
the loaded edge nodes coupled with constraint equations as described in 1550HFigure 7.2. It
was often observed that the constraints used to enforce uniform displacements in 1 (u) at
the loaded edges and the transverse midline restraint in 2 (w=0) were contributing to
solution convergence problems, and therefore an alternative set of boundary conditions
was developed as shown in 1551HFigure 7.8.
1
2
3
Perimeter restrained in 2 (v=0), unloaded edges free to wave
Longitudinal midline restrained in 1 (u=0)
Loaded edge coupled to move together in 1 using rigid body formulation in ABAQUS (all u are equal)
Restraint at center of loaded edge in 3 (w=0)
Restraint at center of loaded edge in 3 (w=0)
Loaded edge coupled to move together in 1 using rigid body formulation in ABAQUS (all u are equal)
Figure 7.8 Stiffened element boundary conditions with rigid body coupling at loaded edges
270
A comparison of the geometric imperfections assumed for the stiffened element
with and without the hole is provided in 1552HFigure 7.9. d1/t=0.34 is used to scale the initial
imperfection field of the plate. This magnitude corresponds to a probability of
occurrence of P(∆<d1)=0.50 (see Section 1553H7.1.1 for details). The peak load of the stiffened
element is sensitive to initial geometric imperfections, and therefore it is important to
consider the same imperfection shape when comparing the load‐displacement responses
of the stiffened element with and without a hole. The imperfection shape is imposed on
the stiffened element with the slotted hole by mapping the buckled mode shape to nodal
coordinates using custom MATLAB code (Mathworks 2006).
fundamental buckling mode mapped to plate with slotted hole
fundamental buckling mode of plate
initial geometric
imperfection fields
Figure 7.9 Initial geometric imperfection field used for the stiffened element with and without a hole
Eight exploratory ABAQUS models are evaluated in this study, each solving the
same nonlinear problem of a stiffened element with a slotted hole compressed uniaxially
until failure as shown in 1554HFigure 7.10. Each model employs a different combination of
ABAQUS solution controls and boundary conditions as summarized in 1555HTable 7.1.
271
Figure 7.10 Deformation at ultimate load of a stiffened element with a hole loaded in compression. The common failure mechanism is material yielding adjacent to the hole followed by plate folding.
Table 7.1 Summary of nonlinear finite element models and associated solution controls
residual limits
damping factor
Io Ir Ic Rnα c
RIKS1 *STATIC, RIKS --- --- --- --- 0.05 0.2 4 (D) 8 (D) 16 (D) 0.005 (D) --- NORIKS2 *STATIC, RIKS --- --- --- --- (D) (D) 4 (D) 8 (D) 16 (D) 0.005 (D) --- NORIKS3 *STATIC, RIKS --- --- --- --- 0.05 0.05 8 16 33 0.005 (D) --- NO
STATIC1 *STATIC 0.01 1 1.00E-20 0.01 --- --- 8 16 33 0.005 (D) --- NOSTATIC2 *STATIC 0.01 1 1.00E-20 0.01 --- --- 8 16 33 0.005 (D) --- YESSTAB1 *STATIC, STABILIZE 0.01 1 1.00E-20 0.01 --- --- 8 16 33 0.005 (D) 0.0162 NOSTAB2 *STATIC, STABILIZE 0.01 1 1.00E-20 0.01 --- --- 8 16 33 0.1 0.0162 NOSTAB3 *STATIC, STABILIZE 0.01 1 1.00E-20 0.01 --- --- 8 16 33 0.005 (D) 0.0162 NO
imposed displacement, rigid body constraintuniform load, rigid body constraint
line search
Loading Type and Boundary Conditions
uniform load, equation constraintuniform load, equation constraintimposed displacement, rigid body constraintuniform load, equation constraintuniform load, equation constraintimposed displacement, rigid body constraint
Solution Controls
initial step size
total time
min step size
max step size
initial step size
max step size
iteration limitsModel ABAQUS Method
*STATIC *STATIC, RIKS
272
1.3.2 171BModified Riks method solution controls
The RIKS1 and RIKS2 finite element models are loaded with a uniformly
distributed load at both ends as shown in 1556HFigure 7.3(a), where equation constraints
couple the loaded edge nodes (see 1557HFigure 7.2). The initial and maximum load step
magnitudes are defined for RIKS1 based on experience gained from the study in Section
1558H7.1.2.1. The RIKS2 model allows ABAQUS to select all load stepping parameters
automatically.
-2 -1.5 -1 -0.5 0 0.5 1 1.5 2-1
-0.8
-0.6
-0.4
-0.2
0
0.2
0.4
0.6
0.8
1
δ/t
λP/P
y,g
RIKS1RIKS2
Initial imperfection shape (scale exaggerated)
Pb
λ
Pb
λ
compression
tension 2
3
δδ
cannot move past peak load
1
Figure 7.11 Load‐displacement curve for the RIKS1 and RIKS2 models showing direction reversal along load path
The load‐displacement responses from the RIKS1 and RIKS2 models are
compared in 1559HFigure 7.11. For both models, ABAQUS does not capture the peak load
and reverts back along the equilibrium path until the plate is loaded to failure in tension!
The ABAQUS Theory Manual states that this type of direction switch is possible when
273
the equilibrium path exhibits very high curvature (ABAQUS 2007a). The ABAQUS
message files (.msg) for these models report that the moment residuals are too high at
the loaded edge nodes and at nodes along the transverse midline of the plate, suggesting
that these boundary conditions are contributing to the convergence difficulties for the
solution.
The RIKS3 model is loaded with imposed displacements at both ends as shown
in 1560HFigure 7.3 (b), where the midline constraint is removed and the loaded edge nodes are
coupled with a rigid body constraint instead of an equation constraint in ABAQUS (see
1561HFigure 7.8). According the ABAQUS Analysis User’s Manual, only the reference node
governing the motion of the rigid body is involved in element level calculations. This
improves computational efficiency and releases the solution algorithm from the force
and moment residual minimization constraints for all nodes in the rigid body except the
reference node.
The solution results from the stiffened elements loaded with consistent nodal
loads (RIKS1, RIKS2) and imposed displacements (RIKS3) are compared in 1562HFigure 7.12.
Before yielding occurs, the three models produce nearly identical load‐displacement
results. As yielding initiates, the RIKS3 model predicts a peak load and post‐peak
response for the stiffened element. This comparison demonstrates that imposed
displacements and rigid body constraints (in contrast to applied loads and equation
constraints) improve the chances for convergence in this case.
274
0 0.2 0.4 0.6 0.8 10
0.2
0.4
0.6
0.8
1
-2 -1.5 -1 -0.5 0 0.5 1 1.5 2-1
-0.8
-0.6
-0.4
-0.2
0
0.2
0.4
0.6
0.8
1
δ/t
λP/P
y,g o
r P/P
y,g
RIKS1, RIKS2RIKS3
0.24 0.29 0.340.35
0.37
0.39
Riks Method finds post-peak path when plate is compressed with imposed displacements
Riks Method retreats back along elastic load path when plate is loaded with consistent nodal loads
δδ
hPλ
hPλ
Figure 7.12 RIKS1 and RIKS2 models experience convergence problems and return along the loading path, the RIKS3 model successfully predicts a peak load and finds a post‐peak load path
1.3.3 172BNewton‐Raphson method
The STATIC1 and STATIC2 models employ the Newton‐Raphson algorithm with
uniform displacements at the loaded edges imposed with equation constraints (see
1563HFigure 7.2). The stepping parameters are chosen to ensure at least 100 increments are
achieved before completion of the simulation. The number of convergence criteria
iterations is also modified by doubling the ABAQUS parameters Io, Ir, and Ic from their
default values (see 1564HTable 7.1). Io represents the number of equilibrium iterations before a
check is performed to ensure that the magnitudes of the moment and force residual
vectors are decreasing. After Io iterations, if the residuals are not decreasing between
two consecutive equilibrium iterations then the length of the increment step is reduced
and the equilibrium search is restarted. Ir represents the number of equilibrium
275
iterations after which the logarithmic rate of convergence check begins. Ic represents the
maximum number of equilibrium iterations within a time increment step. A line search
algorithm is also employed in the STATIC2 model to improve the convergence of the
Newton‐Raphson algorithm when nodal force and moment residuals are large. This
algorithm finds the solution correction vector which minimizes the out‐of‐balance forces
in the structural system (ABAQUS 2007a).
0 0.25 0.5 0.75 1 1.25 1.5 1.750
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1
δ/t
P/P
y,g
STATIC1STATIC2
0.26 0.27 0.280.36
0.365
0.37
δ/t
P/P
y,g
Initial imperfection shape (scale exaggerated)
hP
Displacement control
δδ
STATIC2 (with line search algorithm) finds post-peak equilibrium path before terminating
hP
Figure 7.13 STATIC1 and STATIC2 load‐displacement curves demonstrate convergence difficulties near the peak load.
1565HFigure 7.13 compares the STATIC1 and STATIC2 load‐displacement curves. The
STATIC1 model finds the peak load but then terminates due to moment residual
convergence issues as it attempts to predict the first step of the post‐peak response. The
ABAQUS message (.msg) file for this model states that the moment residuals at nodes
along the loaded edges, along the transverse midline of the plate, and at some nodes
276
near the hole are increasing and convergence is judged unlikely. The solution is
terminated after the automatic time stepping procedure requires a smaller time step
than the minimum set in this model (1x10‐20). The STATIC2 model with the line search
algorithm also finds the peak load of the stiffened element and is able to track onto a
post‐peak equilibrium path before terminating from the same convergence problems
experienced by the STATIC1 model. The success of the line search algorithm in
identifying a post‐peak equilibrium path highlights its potential for solving nonlinear
problems, although a significant increase in computational effort (almost twice the
wallclock time) was also noted.
1.3.4 173BNewton‐Raphson with artificial damping
The STAB1 and STAB2 models solve the stiffened element problem using a
displacement control Newton Raphson algorithm coupled with the automatic artificial
damping discussed in Section 1566H7.1.2.2. The boundary conditions are modified to those
summarized in 1567HFigure 7.8 because of the convergence issues observed with the
constraint equations and transverse midline restraints. As in the case of STATIC1 and
STATIC2, the convergence iteration limits Io, Ir, and Ic are doubled from their default
values. In an attempt to alleviate the moment residual convergence issues from
previous runs, the Newton Raphson parameter Rαn is modified to relax the residual
requirements when the solution approaches the peak load. Rαn is the allowable limit on
the ratio of the largest residual force or moment at a node (rαmax) to the largest change in
force or moment at a node averaged over each time step increment that has been
277
completed (qα). The α superscript indicates that Rαn can be defined for either a
displacement field u or a rotation field Φ. The convergence limit can be written
mathematically as:
ααα qRr n≤max . (7.3)
The default for Rαn of 0.005 is used in STAB1 for both u and Φ fields, whereas in STAB2
Run =0.005 and Rϕn =0.100.
0 0.25 0.5 0.75 1 1.25 1.5 1.750
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1
δ/t
P/P
y,g
STAB1STAB2
0.25 0.3 0.350.3
0.34
0.38
δ/t
P/P
y,g
Highly nonlinear post-peak equilibrium path found with STAB1 and STAB2
Initial imperfection shape (scale exaggerated)
hP
Displacement control
δδ
hP
Figure 7.14 STAB1 and STAB2 load‐displacement curves demonstrate a highly nonlinear post‐peak equilibrium path
The ABAQUS solutions from models STAB1 and STAB2 in 1568HFigure 7.14
demonstrate a highly nonlinear post‐peak equilibrium path. Both models are able to
successfully predict the peak load and then move to a secondary load path. The solution
terminates because the maximum number of Newton‐Raphson iterations is reached.
The modification of the moment residual limit Rϕn from 0.005 to 0.100 did not influence
the solution results.
278
The STAB3 finite element model employs a uniform loading with the Newton‐
Raphson algorithm and artificial damping to determine the nonlinear response of the
stiffened element with a slotted hole. The STAB3 boundary conditions are the same as
those for the STAB1 and STAB2 models, where the plate edges are constrained to move
to together as rigid bodies (see 1569HFigure 7.8). 1570HFigure 7.15 compares the STAB1 and STAB2
(displacement control) to the STAB3 (load control) results and shows that, prior to
yielding, the three models predict the same response. Differences in the load paths are
observed after yielding though, especially in the STAB3 model, which reaches peak load
and then carries this load with zero stiffness over a large deformation range. This
unstable post‐peak behavior results from a complete loss of stiffness as the hole
collapses under load control. The peak loads predicted for the stiffened element by the
displacement control STAB3 model is seven percent higher than the STAB1 and STAB2
load control solutions, demonstrating that the peak load is sensitive to the loading
method (uniform load or uniform displacements) in this case.
279
0 0.2 0.4 0.6 0.8 10
0.2
0.4
0.6
0.8
1
0 0.25 0.5 0.75 1 1.25 1.5 1.750
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1
δ/t
P/P
y,g
STAB1, STAB2STAB3
0.28 0.32 0.360.3
0.35
0.4
δδ
hP
hP
Load control solution demonstrates complete loss of stiffness at peak load (yielding and plate folding at hole)
Higher peak load found with load control compared to displacement control (0.395Py,g versus 0.375 Py,g)
Figure 7.15 The STAB1 and STAB2 models (artificial damping, displacement control) exhibit a sharp drop in load as folding of the plate initiates near the hole. The STAB3 model (artificial damping, load control) finds
the compressive load at which a complete loss of stiffness occurs.
1.4 98BInfluence of a slotted hole on the ultimate strength of a stiffened element (without geometric imperfections)
The solution controls from the previous section resulting in successful
simulations are now implemented to evaluate the influence of a slotted hole on the
ultimate strength and failure mode of a stiffened element. The loading and boundary
conditions, dimensions, material properties, and solution controls are the same as those
used for the STAB2 model described in Section 1571H7.1.3.4 and 1572HTable 7.1. Initial geometric
imperfections are not considered.
280
0 0.2 0.4 0.6 0.8 10
0.2
0.4
0.6
0.8
1
0 0.25 0.5 0.75 1 1.25 1.5 1.750
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1
δ/t
P/P
y,g
Py,net=0.56*Py,g
Pcr,no hole=0.33*Py,g
Pcr,hole=0.30*Py,g
Elastic Buckling
Ultimate Limit State
Figure 7.16 Comparison of ultimate limit state and elastic buckling plate behavior, initial imperfections are not considered in these results
1573HFigure 7.16 demonstrates that the slotted hole reduces the strength of the stiffened
element from 1.0 Py,g to 0.58 Py,g, where Py,g is the resultant compressive force on the
stiffened element to cause yielding calculated with the gross cross‐sectional area of
the plate. The predicted peak load of the stiffened element with the hole is
consistent with the load at yielding of the net section, Py,net=0.56 Py,g. This observation,
that the strength of the stiffened element with the hole is limited to Py,net, highlights
an important consideration in the development of the Direct Strength Method in
1574HChapter 8. The hole also reduces the axial stiffness of the stiffened element in this
case, as demonstrated by the change in slope of the linear portion of the load‐
displacement curve in 1575HFigure 7.16.
281
1.5 99BInfluence of geometric imperfection magnitudes on the ultimate strength of a stiffened element with and without a slotted hole
The ultimate strength of cold‐formed steel members is sensitive to initial
geometric imperfections. In this study the influence of imperfection magnitude on the
ultimate strength of stiffened elements with and without a slotted hole is evaluated. The
loading and boundary conditions, dimensions, material properties, and solution controls
are the same as those used for the STAB2 model discussed in Section 1576H7.1.3.4 and
summarized in 1577HTable 7.1. The imperfection field is assumed as the fundamental elastic
buckling mode pictured in 1578HFigure 7.9. Local buckling (Type 1) imperfection magnitudes
corresponding to P(∆<d1)=0.25, 0.50, 0.75, 0.95, and 0.99 from the CDF in 1579HFigure 7.32 are
considered.
1580HFigure 7.17 and 1581HFigure 7.18 present the load‐displacement results for the stiffened
element without and with the hole and demonstrate that increasing imperfection
magnitudes reduces peak load and change post‐peak response. The elastic stiffness is
also softened, which can be observed by comparing the imperfection results to the linear
slope of the load‐displacement curve without imperfections. This softening results from
the initial out‐of‐plane deformations which engage more of the bending stiffness of the
plate and less of the axial stiffness as the plate is compressed. Out‐of‐plane
deformations (such as initial imperfections) increase the magnitude of the geometric
stiffness matrix which negates the initial elastic stiffness of the undeformed system.
Imperfections are observed to decrease strength but increase ductility of the stiffened
element with and without a hole. The load‐displacement results also highlight that the
282
hole reduces the ductility of the stiffened element, which is consistent with the column
experiment results in 1582HChapter 5.
0 0.25 0.5 0.75 1 1.25 1.5 1.750
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1
δ/t
P/P
y,g
no imperfectionsd1/t=0.14
d1/t=0.34
d1/t=0.66
d1/t=1.35
d1/t=3.85
δ
δDisplacement Control h
P
hP
Figure 7.17 Load‐displacement sensitivity to imperfection magnitude for a plate without a hole
0 0.25 0.5 0.75 1 1.25 1.5 1.750
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1
δ/t
P/P
y,g
no imperfectionsd1/t=0.14
d1/t=0.34
d1/t=0.66
d1/t=1.35
d1/t=3.85
hP
hP
Displacement control
δδ
Figure 7.18 Load‐displacement sensitivity to imperfection magnitude for a plate with a slotted hole
283
1.6 100B Determination of unstiffened element “effective width” using nonlinear finite element modeling
The “effective width” method provides an approximation to the complex non‐
uniform stress distribution in a thin buckled plate under compression. Initially
presented in the 1930’s by von Karman and extended to cold‐formed steel members by
Winter in the 1940’s, the method accounts for the reduction in load‐carrying capacity of
a stiffened element (von Karman et al. 1932; Winter 1947). The inability of the center of
the plate to carry compressive load is caused by out‐of‐plane deformations in the shape
of the fundamental elastic buckling mode. These deformations reduce the axial stiffness,
concentrating the compressive force at the edges of a plate. The ultimate load is reached
when these edge stresses, carried by the “effective width”, exceed the yield stress of the
plate material. The “effective width” concept is the basis of most cold‐formed steel
design codes around the world today.
In this study, a nonlinear finite element model is employed to calculate the
longitudinal stress distribution at failure for a stiffened element with and without a
slotted hole. The distribution of stresses for both cases is compared, and the variation in
effective width along the length of the stiffened element is determined. The stiffened
element is modeled with the same loading and boundary conditions, dimensions,
material properties, and solution controls as those used for the STAB2 model discussed
in Section 1583H7.1.3.4 and described in 1584HTable 7.1. The initial imperfection geometry
corresponds to the fundamental elastic buckling mode of the plate without the hole as
described in 1585HFigure 7.9. d1/t=0.34 is used to scale the initial imperfection field of the
284
plate, which corresponds to a probability of occurrence of P(∆<d1)=0.50 as discussed in
Section 1586H7.1.1. The effective width is calculated by first integrating the longitudinal (S11)
membrane stresses at cross‐sections along the length of the stiffened element and then
dividing the resulting areas by the yield stress of the steel as shown in 1587HFigure 7.19. The
membrane stresses are the longitudinal (S11) stresses that occur at the midplane of the
stiffened element as defined in 1588HFigure 7.20.
h
he/2
he/2membrane stress (S11)
yield stress
calculate area under stress curve (A)
distribute area (A) to edges of plate
A/2
A/2
Figure 7.19 Calculation of “effective width” at a cross‐section along a stiffened element
Plan view of element
+S11 +S11
Elevation view of element
Top
Bottom
MidplaneMembrane stress
Membrane stress
Figure 7.20 Definition of longitudinal (S11) membrane stress
285
1589HFigure 7.21(a) highlights the variation in membrane longitudinal stress (S11)
occurring at the failure load of the stiffened element. The highest stresses accumulate
along the edges of the plate and decrease toward the center of the plate. The largest
edge stresses occur at the crests of the half‐waves where the grey stress contours indicate
yielding of the plate. The corresponding effective width is presented in 1590HFigure 7.21(b).
The maximum effective width of 0.51 he/h occurs at the inflection point between half‐
waves, while the minimum effective width of 0.48 he/h occurs at the wave crests. The
predicted effective width for this plate using Section B2.1 of the AISI specification is 0.50
he/h (AISI‐S100 2007).
(Ave. Crit.: 75%)Mid, (fraction = 0.0)S, S11
-3.300e+01-3.021e+01-2.742e+01-2.462e+01-2.183e+01-1.904e+01-1.625e+01-1.345e+01-1.066e+01-7.871e+00-5.078e+00-2.286e+00+5.059e-01
-3.578e+01
he/2
(a) membrane stress in 1 direction (S11)
Plan view of element
+S11 +S11
Elevation
(b) variation in effective width along plate
h
effective width he/haverage 0.51
standard deviation 0.02max 0.55min 0.48
Figure 7.21 (a) longitudinal membrane stresses and (b) effective width of a stiffened element at failure
The failure mode of the stiffened element with the slotted hole is fundamentally
different than without the hole. The stresses in 1591HFigure 7.22(a) demonstrate that yielding
occurs only at the location of the hole when the peak load is reached. Compressive
stresses are highest at the edge of the plate and then transition to tensile stresses at the
286
face of the hole. The effective width of the yielded portion of the plate in 1592HFigure 7.22(b)
is less than that for the plate without the hole, even with the beneficial tensile stresses at
the face. The average effective width is 0.38 he/h, which is 25 percent less than that of the
stiffened element without the hole. The predicted effective width using Section B2.2 of
the AISI Specification is 0.30 he/h. The effective widths of the stiffened element with and
without a slotted hole are compared together in 1593HFigure 7.23.
(Ave. Crit.: 75%)Mid, (fraction = 0.0)S, S11
-3.300e+01-2.929e+01-2.558e+01-2.187e+01-1.816e+01-1.445e+01-1.074e+01-7.034e+00-3.324e+00+3.852e-01+4.095e+00+7.804e+00+1.151e+01
-3.615e+01
Plan view of element
+S11 +S11
Elevation
(a) membrane stress in 1 direction (S11)
(b) variation in effective width along plate
h
effective width he/haverage 0.38
standard deviation 0.03max 0.41min 0.34
he/2
Figure 7.22 (a) longitudinal membrane stresses and (b) effective width of a stiffened element with a slotted hole at failure
he/2 (plate with hole)he/2 (plate without a hole)
Figure 7.23 Effective width comparison for a plate with and without a slotted hole
The longitudinal stresses (S11) in the top and bottom fibers of the stiffened
element at failure are different from the membrane stresses at the midplane, suggesting
287
that the effective width of a stiffened element actually varies through its thickness.
1594HFigure 7.24 and 1595HFigure 7.25 provide a comparison of this variation for a stiffened
element with and without a slotted hole. It is observed that a plate is more effective on
the surface where the out‐of‐plane deformation causes compression. The effective width
is reduced when tensile and compressive stresses negate each other, as shown in the 2D
plot of extreme fiber and membrane stresses at a representative cross‐section in 1596HFigure
7.26.
Top of plate
Midplane of plate
Bottom of plate
Effective width calculated with longitudinal stresses (S11) at top, midplane, and bottom of the plate
Figure 7.24 Through the thickness variation of effective width of a plate without a hole
Top of plate
Midplane of plate
Bottom of plate
Effective Width calculated with longitudinal stresses (S11) at top, midplane, and bottom of the plate
Figure 7.25 Through the thickness variation of effective width of a plate with a slotted hole
288
-1.5 -1 -0.5 0 0.5 1 1.50
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1
fplate/fy
x/h
top of platemidplane of platebottom of plate
Top of plate is fully effective
Tension and compression stresses counteract each other when calculating effective width at the bottom of the plate
Stress distribution used to calculate code-based effective width
TensionCompression
Longitudinal (S11) stress variation across width of plate
A
A
SECTION A-A
Figure 7.26 Through thickness variation in longitudinal (S11) stresses in a plate at failure
7.2 47BNonlinear finite element modeling of columns with holes
A more extensive study of ABAQUS nonlinear finite element capabilities of cold‐
formed steel columns with holes is now presented. Simulation to collapse of the 24
column experiments described in 1597HChapter 5 is performed, considering solution
sensitivity to specific modeling parameters including initial imperfections, residual
stresses and the cold‐work of forming, nonlinear material modeling, and column
boundary conditions. A modeling protocol is developed which produces results
consistent with column experiments. This modeling tool is employed to explore the
289
validity of proposed Direct Strength Method equations for cold‐formed steel members
with holes presented in 1598HChapter 8.
2.1 101BModeling protocol development
2.1.1 174BModel dimensions and finite element meshing
The collapse behavior of the 24 column specimens is simulated with the general
purpose finite element program ABAQUS (ABAQUS 2007a). All columns are modeled
with S9R5 reduced integration nine‐node thin shell elements (see Section 1599H2.1 for details
on the S9R5 element). The finite element mesh for each specimen is created with custom
Matlab code developed by the author (see 1600HAppendix A); the mesh is consistent with
S9R5 meshing guidelines summarized in Section 1601H2.4. The centerline C‐section
dimensions input into ABAQUS are calculated using the out‐to‐out dimensions of each
column specimen provided in 1602HTable 5.3. The cross‐section corner angles are assumed as
right angles (even though they were measured to be off of 90 degrees, see 1603HTable 5.4)
since the distortional imperfection magnitudes obtained in Section 1604H7.2.1.5 are derived
based on a nominal cross‐section with 90 degree corners*. The average base metal
thickness for each specimen (i.e., the average of tbare,w, tbare,f1, and tbare,f2 from 1605HTable 5.5) and
column length L from 1606HTable 5.6 are used to construct the ABAQUS models, as are the
location of the slotted web holes relative to the centerline of the web provided in 1607HTable
5.8.
* The measured flange‐web and web‐lip angles were not considered because of initial difficulties matching the experiment results to the simulations. To resolve these difficulties, a simplified model with nominal dimensions was implemented. (Modeling with plasticity at the proportional limit was found to be the true cause of the discrepancies, see Section 1608H7.2.1.4.) Consideration of the actual cross‐section dimensions, including the flared web‐flange corners measured in the experiments, is warranted and is an important point of future study.
290
2.1.2 175BBoundary conditions and application of loading
The specimen boundary conditions in ABAQUS are defined to simulate the
experiment boundary conditions as shown in 1609HFigure 7.27. The nodes on the loaded
column face are coupled together in the direction of loading (1 direction) with an
ABAQUS “pinned” rigid body constraint. This constraint ensures that all nodes on the
loaded face of the column translate together, while the rotational degrees of freedom
remain independent (as in the case of platen bearing). A total imposed displacement of
0.20 inches is applied to the reference node of the ABAQUS rigid body over a series of
steps (see Section 1610H7.2.1.3) to simulate the displacement control loading applied by the
bottom platen during the experiment. Friction‐contact boundary conditions were also
evaluated in ABAQUS as described in 1611HAppendix J although their influence on the
ultimate strength of the column specimens was determined to be minimal.
1
2
3
ABAQUS “pinned “ rigid body reference node constrained in 2 to 6 directions, ensures that all nodes on loaded surface move together in 1 direction
Nodes bearing on top platen constrained in 1, 2 and 3
45
6Apply uniform displacement in ABAQUS to simulate displacement control of experiments
δ
Figure 7.27 ABAQUS boundary conditions simulating column experiments
291
2.1.3 176BNonlinear solution method
The modified Riks method (*STATIC, RIKS) is employed as the solution algorithm in
this study. The preliminary nonlinear finite element studies on stiffened elements
demonstrated that the modified Riks method was able to capture the complete load‐
deformation response when imposed displacements are used to load the member (see
1612HFigure 7.12). ABAQUS automatic time stepping was enabled, with the suggested initial
step size set to 0.005, the maximum step size limited to 0.01, and the maximum number
of increments equal to 300 all input by the user.
2.1.4 177BMaterial modeling
Steel yielding and plasticity is simulated in ABAQUS using a classical metal
plasticity approach with isotropic hardening. A Mises yield surface is defined with the
true stress and true plastic strain obtained from uniaxial tensile coupon tests for each
specimen. Three stress‐strain curves (west flange, east flange, and web) were obtained
for each specimen (see Section 1613H5.2.5). The experimentally obtained engineering stress‐
strain curves are converted to true stress and strain and then averaged point‐by‐point to
produce a yield stress, proportional limit, and true stress‐strain curve for each specimen.
(The true plastic strains and associated stresses are input into ABAQUS with the
*PLASTIC command.) For Mises stresses below the yield stress, linear elastic material
behavior is assumed where E=29500 ksi and ν=0.3.
292
Preliminary nonlinear modeling efforts for this study determined that including
plastic strains starting at the proportional limit resulted in ABAQUS simulation
predictions that were as much as 25% lower than the column tested strengths. An
example of the average true stress‐strain curves with plasticity starting at the
proportional limit are provided in 1614HFigure 7.28a for specimen 362‐1‐24‐NH and 1615HFigure
7.29a for specimen 600‐1‐24‐NH (average true stress‐strain curves are provided for all 24
column specimens in 1616HAppendix H). Other researchers have obtained simulation results
consistent with experiments by assuming that plastic strains initiate only after the yield
stress (determined with the 0.2% offset method) is reached (Schafer 1997; Yu 2005;
Schafer et al. 2006) ‐ a material modeling approach that proved to be successful at
predicting the column experiment peak loads for this study also. The true stress‐strain
curves in 1617HAppendix H were therefore modified to ensure that plasticity initiates in
ABAQUS only after the yield stress is reached for the gradually yielding stress‐strain
curves (362S162‐33 specimens, see 1618HFigure 7.28b) and at the initiation of the yield plateau
for the sharp‐yielding curves (600S162‐33 specimens, see 1619HFigure 7.29b). 1620HFigure 7.30
demonstrates the disparity between the experiment and FE simulation load‐deformation
response for column specimen 600‐1‐24‐NH when plasticity initiates at the proportional
limit. The reasons for this disparity are unclear and warrant future study. Additional
research work is planned to study the details of metal plasticity in ABAQUS by loading
a single finite element to failure in tension.
293
0 0.02 0.04 0.06 0.08 0.1 0.120
10
20
30
40
50
60
70
80
90
100
true plastic strain
true
stre
ss, k
si
0 0.02 0.04 0.06 0.08 0.1 0.120
10
20
30
40
50
60
70
80
90
100
true plastic strain
true
stre
ss, k
si
true plastic strain true stress
ksi0 55.1
0.003 60.30.008 64.90.013 68.40.023 74.00.033 78.10.043 81.30.053 83.80.063 86.2
true plastic strain true stress
ksi0 33.1
0.001 46.10.002 51.90.007 60.30.012 64.90.017 68.40.027 74.00.037 78.10.047 81.30.057 83.80.067 86.2
(a) (b)
Figure 7.28 ABAQUS plastic strain curve for specimen 362‐1‐24‐NH assuming (a) plasticity initiates at the proportional limit and (b) plasticity initiates at 0.2% offset yield stress
0 0.02 0.04 0.06 0.08 0.1 0.120
10
20
30
40
50
60
70
80
90
100
true plastic strain
true
stre
ss, k
si
0 0.02 0.04 0.06 0.08 0.1 0.120
10
20
30
40
50
60
70
80
90
100
true plastic strain
true
stre
ss, k
si
true plastic strain true stress
ksi0 37.5
0.001 54.70.002 58.30.007 60.00.012 61.50.017 64.00.027 70.20.037 74.40.047 77.50.057 80.00.067 81.90.077 83.50.087 84.90.097 86.10.107 87.2
(a) (b)
true plastic strain true stress
ksi0 58.3
0.005 60.00.01 61.5
0.015 64.00.025 70.20.035 74.40.045 77.50.055 80.00.065 81.90.075 83.50.085 84.90.095 86.10.105 87.2
Figure 7.29 ABAQUS plastic strain curve for specimen 600‐1‐24‐NH assuming (a) plasticity initiates at the proportional limit and (b) plasticity initiates at the beginning of the yield plateau (refer to 1621HAppendix H for
the details on the development of this curve).
294
0 0.02 0.04 0.06 0.08 0.1 0.12 0.14 0.16 0.18 0.20
2
4
6
8
10
12
14
axial displacement, in.
axia
l loa
d, k
ips
plasticity at proportional limitplasticity at yield stressexperiment
Figure 7.30 Influence of ABAQUS material model on the load‐deformation response of specimen 600‐1‐24‐
NH (work this figure with 1622HFigure 7.29)
2.1.5 178BInitial geometric imperfections
The ultimate strength and failure mechanisms of cold‐formed steel columns are
sensitive to initial geometric imperfections, as demonstrated in the preliminary studies
on stiffened elements in Section 1623H7.1.5. In this study, the sympathetic local (L) and
distortional (D) elastic buckling modes are obtained with eigenbuckling analyses for
each column specimen and imposed on the nominal geometry in each finite element
model. (An ABAQUS .fil file is created for each eigenbuckling analysis which is then
called from the nonlinear .inp file with the *IMPERFECTION command). The boundary
conditions at both specimen ends are assumed to be warping free when obtaining the
imperfection shapes (see 1624HFigure 4.2 for definition) to ensure consistency with CUFSM
boundary conditions. Specimens with and without holes are modeled with the same
elastic buckling imperfection shapes by filling the holes with additional finite elements
295
as shown in 1625HFigure 7.31. This procedure ensures that the load‐displacement behavior of
both the hole and no hole specimens are compared on equivalent basis (both will have
the no hole L and D imperfection shapes). Filling in the holes is necessary (instead of
eliminating them completely) because it preserves the nodal numbering and geometry
of the specimens with holes, making it convenient to superimpose the L and D modes
onto the initial nodal geometry in ABAQUS.
Hole is filled in with S9R5 elements to produce no hole local buckling shape
Type1 imperfection (L) Type 2 imperfection (D)
Figure 7.31 Slotted holes are filled with S9R5 elements to obtain no hole imperfection shapes
The magnitudes of the L and D imperfections are determined with the same
probabilistic treatment used for the stiffened element studies in Section 1626H7.1 (Schafer and
Peköz 1998). Finite element simulations with L and D imperfection magnitudes
corresponding to the 25th and 75th percentiles of the CDF in 1627HFigure 7.32 are performed for
each specimen to obtain a range of simulated load‐displacement responses to compare
to experiment results. FE simulations are also performed using the L and D elastic
296
buckling mode shapes and imperfection magnitudes measured directly from the column
specimens. In this case the local imperfection magnitude for each specimen is taken as
the maximum deviation from the average web elevation as reported in 1628HTable 5.9. The
distortional imperfection magnitude for each specimen is determined by finding the
largest measured angular deviation from 90 degrees along each specimen and
calculating the associated flange‐lip displacement as shown in 1629HFigure 7.33. The Type 1
imperfection magnitudes measured in the experiments are often 2 to 3 times larger than
the 75th percentile CDF magnitudes as shown in 1630HTable 7.2. The Type 2 imperfection
magnitudes for the 362S162‐33 specimens also are 2 to 3 times larger than the 75th
percentile CDF magnitudes, primarily because these specimens tended to open up at the
sawn ends (i.e., flange‐web angles increased above 90 degrees) when they were saw‐cut
from full stud lengths. Other researchers have studied this observed change in cross‐
section after saw‐cutting (Wang et al. 2006). The 600S162‐33 specimens were less
sensitive to this saw‐cutting effect, resulting in measured distortional imperfection
magnitudes consistent with the 75th percentile of the imperfection CDF.
CDF of Maximum Imperfection
Type 1 Type 2
(L) (D)
d
d
0.00
0.20
0.40
0.60
0.80
1.00
0.0 1.0 2.0 3.0 4.0 5.0 6.0 7.0
d/t
Prob
abili
ty (X
< x
)
Type 1Type 2
Type 1 Type 2P(X < x) d/t d/t
0.25 0.14 0.640.50 0.34 0.940.75 0.66 1.550.95 1.35 3.440.99 3.87 4.47
μ 0.50 1.29σ 0.66 1.07
0.00
0.20
0.40
0.60
0.80
1.00
0.0 1.0 2.0 3.0 4.0 5.0 6.0 7.0
d/t
Prob
abili
ty (X
< x
)
Type 1Type 2
Type 1 Type 2P(X < x) d/t d/t
0.25 0.14 0.640.50 0.34 0.940.75 0.66 1.550.95 1.35 3.440.99 3.87 4.47
μ 0.50 1.29σ 0.66 1.07
297
Figure 7.32 L and D imperfection magnitudes described with a CDF (Schafer and Peköz 1998)
Nominal cross-section
Deviation from reference cross section (measured at X=6, 18 in. for the the short specimens and X=12, 18, 30 and 36 in. for the 48 in. long specimens).
B1 B2
θ1 θ2
( )iiBD θsinmax(= where i=1 or 2
90 degrees 90 degrees
Figure 7.33 Method for measuring distortional imperfection magnitudes from experiments
Table 7.2 Local and distortional imperfection magnitudes Specimen
25% CDF 75% CDF Measured 25% CDF 75% CDF Measured362-1-24-NH 0.005 0.025 0.038 0.025 0.060 0.200362-2-24-NH 0.005 0.025 0.054 0.025 0.060 0.174362-3-24-NH 0.005 0.025 0.036 0.025 0.060 0.205362-1-24-H 0.005 0.026 0.052 0.025 0.061 0.195362-2-24-H 0.005 0.025 0.058 0.025 0.059 0.159362-3-24-H 0.006 0.026 0.044 0.025 0.061 0.140362-1-48-NH 0.005 0.026 0.071 0.025 0.061 0.168362-2-48-NH 0.006 0.026 0.080 0.025 0.061 0.163362-3-48-NH 0.005 0.026 0.057 0.025 0.060 0.184362-1-48-H 0.005 0.026 0.084 0.025 0.061 0.161362-2-48-H 0.005 0.026 0.066 0.025 0.061 0.174362-3-48-H 0.006 0.026 0.050 0.026 0.062 0.156600-1-24-NH 0.006 0.029 0.061 0.028 0.068 0.106600-2-24-NH 0.006 0.029 0.075 0.028 0.068 0.114600-3-24-NH 0.006 0.029 0.096 0.028 0.068 0.114600-1-24-H 0.006 0.028 0.062 0.027 0.065 0.064600-2-24-H 0.006 0.027 0.089 0.026 0.064 0.124600-3-24-H 0.006 0.028 0.087 0.028 0.067 0.102600-1-48-NH 0.006 0.029 0.071 0.028 0.067 0.090600-2-48-NH 0.006 0.028 0.077 0.028 0.067 0.077600-3-48-NH 0.006 0.029 0.073 0.028 0.067 0.074600-1-48-H 0.006 0.028 0.095 0.027 0.066 0.084600-2-48-H 0.006 0.028 0.049 0.027 0.067 0.062600-3-48-H 0.006 0.028 0.068 0.028 0.067 0.099
Type 2 Imperfection Magnitude (D)Type 1 Imperfection Magnitude (L)
The initial out‐of‐straightness of each column specimen was measured in the MTS
machine under a small preload before the start of each test. This global imperfection is
also superimposed on the nodal geometry for each specimen finite element model as
298
shown in 1631HFigure 7.34. The magnitude of the global imperfection, Δg, is provided in 1632HTable
7.3.
aa
Section a-aSpecimen COG (Typ.)
Δg (+ shown)
Loading Line
Figure 7.34 Definition of out‐of‐straightness imperfections implemented in ABAQUS
Table 7.3 Out‐of‐straightness imperfection magnitudes Δg
in.362-1-24-NH -0.024362-2-24-NH 0.004362-3-24-NH 0.038362-1-24-H -0.012362-2-24-H 0.034362-3-24-H -0.023362-1-48-NH 0.047362-2-48-NH -0.028362-3-48-NH 0.012362-1-48-H 0.066362-2-48-H 0.013362-3-48-H -0.003600-1-24-NH -0.063600-2-24-NH -0.141600-3-24-NH 0.063600-1-24-H -0.078600-2-24-H 0.076600-3-24-H 0.069600-1-48-NH -0.036600-2-48-NH -0.087600-3-48-NH -0.049600-1-48-H -0.098600-2-48-H 0.072600-3-48-H 0.020
Specimen
299
2.1.6 179BResidual stresses and equivalent plastic strains
1633HChapter 6 describes a general method for predicting the through thickness residual
stresses and strains in cold‐formed steel members which can then be readily input into
ABAQUS. The prediction method assumes that residual stresses and plastic strains
occur over the full cross‐section from coiling, uncoiling, and flattening of the sheet coil.
The coiling residual stresses are largest when the sheet thickness t is large (>0.068 in.)
and the yield stress is low (<40 ksi). The predicted coiling, uncoiling, and flattening
residual stresses (and plastic strains) are zero in this study because the column
specimens have a relatively low sheet thickness (~0.040 in.) and high yield stress (~ 60
ksi).
Residual stresses and plastic strains from the roll‐forming of the cross‐section are
considered in this study. These stresses are applied in ABAQUS with the element local
coordinate system shown in 1634HFigure 7.35 starting from section point 1 (SNEG). The
transverse residual stress distribution (2‐direction) is provided in 1635HFigure 7.36 and the
longitudinal distribution (1‐direction) in 1636HFigure 7.37 as a function of yield stress σyield (σyield
is listed in 1637HTable 5.13 for each specimen). Plastic strains are input into ABAQUS in von
Mises space and therefore only plastic strain magnitudes are required, not a specific
direction. The highest initial strains occur at the inner and outer surfaces of the corners
as shown in 1638HFigure 7.38. εp is approximated using the procedure outlined in 1639HFigure 6.17.
300
SNEG
2
1SPOS
Element normal
Figure 7.35 ABAQUS element local coordinate system for use with residual stress definitions
SNEG
SPOS
-0.50σyield
+σyield -σyield
+0.50σyield
2
Figure 7.36 Transverse residual stress distribution applied at the corners of the cross‐section
SNEG
SPOS
+0.05σyield
1
-0.05σyield
-0.50σyield
+0.50σyield
Figure 7.37 Longitudinal residual stress distribution applied at the corners of the cross‐section
301
εp
εpSNEG
SPOS
Figure 7.38 Equivalent plastic strain distribution at the corners of the cross‐section
The transverse residual stress distribution has the special property that it is self‐
equilibrating for both moment and axial force, i.e. the total force and moment through
the thickness is zero. This self‐equilibrating characteristic ensures that no deformation
(or redistribution of stress) will occur in ABAQUS in the initial state. The longitudinal
stress distribution is self‐equilibrating for axial force but not for moment. The
deformations associated with this out‐of‐balance bending moment are infinitesimal and
very small redistributions in stress are observed (±0.1 ksi) in the initial state.
The number of element section points through the thickness dictates the accuracy of
the residual stress distribution. If only a small number of section points are used, the
discontinuity in stress at the middle thickness cannot be modeled accurately and
excessive transverse deformations of the cross‐section will occur. 1640HFigure 7.39
demonstrates the decrease in unbalanced through‐thickness transverse moment, MUB, as
the number of section points increase (sheet thickness is assumed as t=0.040 in. and
σyield=60 ksi). As the number of section points decrease, the residual stress approaches
0.50My, where My is the yield moment of the sheet steel per unit width defined as:
302
yieldy
tM σ6
2
= (7.4)
55 section points are used in the specimen finite element models for this study as a
compromise between model accuracy and computational cost. ABAQUS limits the
maximum number of section points to 250 for the S9R5 element (ABAQUS 2007b).
0 20 40 60 80 1000
0.05
0.1
0.15
0.2
0.25
0.3
0.35
0.4
0.45
0.5
# of element through-thickness section points
MU
B/My
Figure 7.39 Influence of section points on the unbalanced moment (accuracy) of the transverse residual stress distribution as implemented in ABAQUS
2.2 102BModeling protocol validation
2.2.1 180BUltimate strength and failure mechanisms
The nonlinear finite element protocol presented in Section 1641H7.2.1 is demonstrated to
be a viable, conservative predictor of peak load when compared to the experiment
results in 1642HTable 7.4. The mean of the experimental peak load Ptest to ABAQUS peak load
303
PABAQUS ratios are 1.03 (25th percentile imperfection CDF), 1.05 (75th percentile imperfection
CDF), and 1.11 (measured imperfections). In a few cases (and always with specimens
with holes), ABAQUS was not able to obtain the peak load, either because the modified
Riks solution algorithm reversed the direction of the applied load (similar to that
observed in 1643HFigure 7.12 for stiffened elements) or because the ABAQUS could not find
equilibrium and terminated the simulation. As imperfection magnitudes increased, the
modified Riks solution algorithm was more successful at reaching peak load. This trend
is hypothesized to occur because for small imperfection magnitudes a specific
deformation pattern is not established and many equilibrium paths exist near peak load,
whereas for larger imperfection magnitudes a dominate deformation shape and
equilibrium path are defined early in the simulation. Nonlinear FE load‐displacement
behavior is provided for a representative sample of specimens in 1644HFigure 7.40 to 1645HFigure
7.47, including the load‐displacement curves and deformed shape at collapse (compare
these simulated shapes to the pictures of experiments in 1646HAppendix F). FE simulation
load‐displacement curves are provided for all specimens in 1647HAppendix I.
Table 7.4 Comparison of nonlinear FE simulation peak loads to experiments
304
PABAQUS Ptest/PABAQUS PABAQUS Ptest/PABAQUS PABAQUS Ptest/PABAQUS
kips kips kips kips362-1-24-NH 10.48 10.26 1.02 9.88 1.06 8.72 1.20362-2-24-NH 10.51 10.13 1.04 9.70 1.08 8.82 1.19362-3-24-NH 10.15 10.21 0.99 9.85 1.03 8.69 1.17362-1-24-H 10.00 9.22 1.09 9.08 1.10 8.48 1.18362-2-24-H 10.38 8.83 1.18 8.70 1.19 8.27 1.26362-3-24-H 9.94 9.19 1.08 9.11 1.09 8.78 1.13362-1-48-NH 9.09 9.48 0.96 9.34 0.97 7.76 1.17362-2-48-NH 9.49 9.40 1.01 9.27 1.02 8.36 1.14362-3-48-NH 9.48 9.26 1.02 8.89 1.07 7.44 1.28362-1-48-H 8.95 8.97 1.00 8.73 1.02 8.30 1.08362-2-48-H 9.18 8.91 1.03 8.63 1.06 8.26 1.11362-3-48-H 9.37 8.58 1.09 DNC --- ED ---600-1-24-NH 11.93 12.14 0.98 12.03 0.99 11.83 1.01600-2-24-NH 11.95 12.10 0.99 12.01 1.00 11.74 1.02600-3-24-NH 12.24 12.10 1.01 11.99 1.02 11.64 1.05600-1-24-H 12.14 DNC --- 11.63 1.04 11.45 1.06600-2-24-H 11.62 11.10 1.05 11.08 1.05 10.82 1.07600-3-24-H 11.79 DNC --- 11.76 1.00 11.49 1.03600-1-48-NH 11.15 11.27 0.99 11.14 1.00 11.32 0.98600-2-48-NH 11.44 11.27 1.02 11.39 1.00 11.30 1.01600-3-48-NH 11.29 11.37 0.99 11.18 1.01 11.04 1.02600-1-48-H 11.16 DNC --- 10.22 1.09 ED ---600-2-48-H 11.70 DNC --- DNC --- 10.17 1.15600-3-48-H 11.16 DNC --- 10.35 1.078 10.30 1.08Average 1.03 1.05 1.11Standard deviation 0.05 0.05 0.08DNC Did Not Complete, Abaqus terminated before finding the peak loadED Excessive distortion - Abaqus error, imperfection magnitude causes element distortional
SpecimenMeasured imperfections25th percentile
imperfection CDF75th percentile
imperfection CDFPtest
The initial elastic slope of the 25% CDF and 75% CDF FE load‐displacement curves
are consistent with experimental results as shown in 1648HFigure 7.40 to 1649HFigure 7.47,
demonstrating that the elastic material modeling assumptions and specimen dimensions
are consistent with the experiments. The initial slope of the load‐displacement curve is
also sensitive to imperfection magnitudes, and therefore the similarities between
experiment and the FE results confirm the assumption that the 25th and 75th percentile
imperfection magnitudes in the FE simulations produce physically realistic results. This
is contrary to the FE simulations with measured imperfections for the 362S162‐33
specimens (for example, see 1650HFigure 7.40), where the initial load‐displacement slope and
peak load are 15% to 30% less than the experimental results (see 1651HTable 7.4 and 1652HFigure
7.40). The FE simulations for the 600S162‐33 specimens are much less sensitive to
imperfection magnitudes (for example, see 1653HFigure 7.44). The maximum difference in test
to predicted ratio between the three imperfection levels (25% CDF, 75% CDF, and
measured) in 1654HTable 7.4 for the 600S162‐33 specimens is 3%.
305
The post‐peak ductility of the column specimens is often underpredicted in the
ABAQUS nonlinear finite element models. The collapse mechanism of a column dictates
its ductility and in some cases its peak load. For example, outward distortional buckling
has been shown to produce lower column strengths than inward distortional buckling
(Silvestre and Camotim 2005). This observation could explain why the FE simulations of
the 362S162‐33 specimens with holes (which exhibit outward distortional buckling) have
a lower peak load and ductility than the experiment results (all three specimens exhibit
inward distortional buckling, see 1655HAppendix F). Another factor influencing column
ductility may be the ABAQUS material modeling effect discussed in Section 1656H7.2.1.4.
When plasticity is considered at the proportional limit (see 1657HFigure 7.30) the peak of the
load‐displacement curve is flattened which is more consistent with experiment results.
These hypotheses motivate important future work to better understand metal plasticity
and material modeling in ABAQUS and also the influence of imperfection shapes on FE
column ductility and strength predictions.
306
75% Imperfection CDF0 0.02 0.04 0.06 0.08 0.1 0.12 0.14 0.16 0.18 0.20
2
4
6
8
10
12
14
axial displacement, in.
axia
l loa
d, k
ips
ExperimentFE, 25% Imperfection CDFFE, 75% imperfection CDFFE, measured imperfections
Figure 7.40 Load‐displacement response of specimen 362‐1‐24‐NH
75% Imperfection CDF0 0.02 0.04 0.06 0.08 0.1 0.12 0.14 0.16 0.18 0.2
0
2
4
6
8
10
12
14
axial displacement, in.
axia
l loa
d, k
ips
ExperimentFE, 25% Imperfection CDFFE, 75% imperfection CDFFE, measured imperfections
Figure 7.41 Load‐displacement response of specimen 362‐1‐24‐H
307
75% Imperfection CDF0 0.02 0.04 0.06 0.08 0.1 0.12 0.14 0.16 0.18 0.20
2
4
6
8
10
12
14
axial displacement, in.
axia
l loa
d, k
ips
ExperimentFE, 25% Imperfection CDFFE, 75% imperfection CDFFE, measured imperfections
Figure 7.42 Load‐displacement response of specimen 362‐1‐48‐NH
75% Imperfection CDF0 0.02 0.04 0.06 0.08 0.1 0.12 0.14 0.16 0.18 0.2
0
2
4
6
8
10
12
14
axial displacement, in.
axia
l loa
d, k
ips
ExperimentFE, 25% Imperfection CDFFE, 75% imperfection CDFFE, measured imperfections
Figure 7.43 Load‐displacement response of specimen 362‐1‐48‐H
308
75% Imperfection CDF0 0.02 0.04 0.06 0.08 0.1 0.12 0.14 0.16 0.18 0.2
0
2
4
6
8
10
12
14
axial displacement, in.
axia
l loa
d, k
ips
ExperimentFE, 25% Imperfection CDFFE, 75% imperfection CDFFE, measured imperfections
Figure 7.44 Load‐displacement response of specimen 600‐1‐24‐NH
75% Imperfection CDF0 0.02 0.04 0.06 0.08 0.1 0.12 0.14 0.16 0.18 0.2
0
2
4
6
8
10
12
14
axial displacement, in.
axia
l loa
d, k
ips
ExperimentFE, 25% Imperfection CDFFE, 75% imperfection CDFFE, measured imperfections
Figure 7.45 Load‐displacement response of specimen 600‐2‐24‐H
309
75% Imperfection CDF0 0.02 0.04 0.06 0.08 0.1 0.12 0.14 0.16 0.18 0.2
0
2
4
6
8
10
12
14
axial displacement, in.
axia
l loa
d, k
ips
ExperimentFE, 25% Imperfection CDFFE, 75% imperfection CDFFE, measured imperfections
Figure 7.46 Load‐displacement response of specimen 600‐1‐48‐NH
75% Imperfection CDF0 0.02 0.04 0.06 0.08 0.1 0.12 0.14 0.16 0.18 0.2
0
2
4
6
8
10
12
14
axial displacement, in.
axia
l loa
d, k
ips
ExperimentFE, 25% Imperfection CDFFE, 75% imperfection CDFFE, measured imperfections
Figure 7.47 Load‐displacement response of specimen 600‐3‐48‐H
310
2.2.2 181BInfluence of residual stresses and initial plastic strains
Residual stresses (RS) and initial plastic strains (PS) from the manufacturing process
are approximated with the prediction method in 1658HChapter 6 and then input into
ABAQUS as discussed in Section 1659H7.2.1.6. 1660HFigure 7.48 highlights their effect on the load‐
deformation response of specimen 600‐1‐24‐NH. A small increase in peak load
(approximately 2%) is observed when just initial plastic strains are considered at the
corners, which simulates the increase in apparent yield stress from strain hardening.
The increase in strength is minimal because the influence of the stiffened corners is offset
by the large proportion of unformed steel (i.e., flats) in the cross‐section. The transverse
and longitudinal residual stresses created by cold‐bending of the cross‐section also have
a minimal impact on the load‐deformation response for this specimen, primarily because
the plastic strains at the corners are high (εp is predicted to be large as 0.20 at the corner
outer fibers) which increases the apparent yield stress in ABAQUS and prevents a loss in
stiffness at the corners, even with the presence of the through‐thickness residual stresses
in the column. Similar load‐displacement trends are also observed for specimen 362‐1‐
24‐NH as shown in 1661HFigure 7.49.
Residual stresses and plastic strains are expected to have a larger influence on the
ultimate strength of members with cross‐sections made from thicker sheet steel, since
coiling and uncoiling of the sheet steel will impart residual stresses and plastic strains
around the entire cross‐sections (see 1662HFigure 6.19). Future research is planned to study
the influence of through‐thickness residual stresses and plastic strains on yielding
311
patterns and failure modes of cold‐formed steel members. The ABAQUS metal
plasticity model with isotropic versus kinematic hardening also needs further study to
determine if one is better than the other when considering the influence of residual
stresses and strains (see Section 1663H6.7 for a more detailed discussion).
0 0.02 0.04 0.06 0.08 0.1 0.12 0.14 0.16 0.18 0.20
2
4
6
8
10
12
14
axial displacement, in.
axia
l loa
d, k
ips
without RS or PSwith RS and PSexperiment
Figure 7.48 Influence of residual stresses (RS) and plastic strains (PS) on the FE load‐displacement response of specimen 600‐1‐24‐NH
312
0 0.02 0.04 0.06 0.08 0.1 0.12 0.14 0.16 0.18 0.20
2
4
6
8
10
12
14
axial displacement, in.
axia
l loa
d, k
ips
without RS or PSwith RS and PSexperiment
Figure 7.49 Influence of residual stresses (RS) and plastic strains (PS) on the FE load‐displacement response
of specimen 362‐1‐24‐NH.
313
Chapter 8 7BThe Direct Strength Method for cold-formed steel members with holes
The nonlinear finite element capability developed in 1664HChapter 7 is now employed to
evaluate proposed Direct Strength Method (DSM) formulations for cold‐formed steel
members with holes. Several hundred cold‐formed steel columns and beams with
standard SSMA structural stud cross‐sections (SSMA 2001) and with varying web hole
sizes, shapes, and spacings are simulated to collapse in ABAQUS. The elastic buckling
properties of these members (Pcrl, Pcrd, and Pcre for columns and Mcrl, Mcrd, and Mcre for
beams), including the presence of the holes, are approximated with the CUFSM elastic
buckling prediction methods developed in 1665HChapter 4. The corresponding ultimate
strengths (obtained from the ABAQUS simulations) are merged with the elastic buckling
information into a simulated experiments database which is utilized to identify potential
modifications to the existing DSM local, distortional, and global failure prediction
curves. Specific DSM options are proposed from these comparisons, which are then
314
compared to the experiment elastic buckling and tested strength databases in 1666HChapter 4
to formalize the final proposed DSM recommendations for cold‐formed steel members
with holes.
8.1 48BDSM for columns with holes
1.1 103BDatabase of simulated column experiments
Simulated experiments were conducted on 211 C‐section columns with evenly‐
spaced slotted or circular web holes in ABAQUS. Column lengths and cross‐sections
were specifically selected with custom Matlab code employing the existing DSM design
curves to identify columns predisposed to local, distortional, and global buckling type
failures. The cross‐sections were chosen from a catalog of 99 industry standard C‐
sections published by the Steel Stud Manufacturers Association (SSMA 2001). The
nominal out‐to‐out dimensions provided in the SSMA catalog were converted to
centerline dimensions and then constructed in ABAQUS with the meshing procedure
described in Section 1667H7.2.1.1. Evenly spaced circular or slotted web holes were placed in
the columns with hole spacing S (defined in 1668HFigure 3.2) varying between 12 and 22
inches. The holes were centered transversely in the web and their depth, hhole, was varied
such that the ratio of the net cross‐sectional area, Anet, to the gross cross‐sectional area, Ag,
ranged between 0.60 and 1.0.
The ABAQUS boundary conditions and application of loading, described in 1669HFigure
8.1, are implemented to be consistent with CUFSM, i.e. pinned‐pinned and free‐to‐warp
with a uniform stress applied at the member ends. These boundary conditions were
315
specifically chosen to permit the use of CUFSM simplified elastic buckling methods
when predicting the elastic buckling behavior of columns with holes. (If pinned‐pinned
warping‐fixed end conditions or fixed‐fixed end conditions were used the elastic
buckling predictions would have required modifications factors, see Eq. 1670H(4.8) for an
example). CUFSM boundary conditions represent a lower bound on member strength
are therefore considered conservative in design. Consistent nodal loads are applied to
simulate the uniform compressive stress at the column ends (see Section 1671H7.2.1.2 for
information on S9R5 consistent nodal loads). The loads (a reference load of 1 kip was
applied at each end in ABAQUS) are distributed over the first two sets of cross‐section
nodes to avoid localized failures at the loaded edges.
1
2
3
End cross-section nodes restrained in 2 and 3
45
6
End cross-section nodes restrained in 2 and 3
Node centered in flange at longitudinal midline restrained in 1 (to prevent rigid body motion)
Consistent nodal loads applied over two sets of cross-section nodes to avoid edge failures
Figure 8.1 ABAQUS simulated column experiments boundary conditions and application of loading
The ABAQUS simulations were performed with the modified Riks nonlinear
solution algorithm. Automatic time stepping was enabled with a suggested initial arc
316
length step of 0.25 (the Riks method increments in units of energy, in this case kip∙in.), a
maximum step size of 0.75, and the maximum number of solution increments set at 300.
Metal plasticity was simulated with the material modeling procedure described in
Section 1672H7.2.1.4. The plastic true stress‐strain curve for specimen 362‐1‐48‐H in 1673HAppendix
H was assumed for all column models (but modified so that plasticity starts at the yield
stress, see Section 1674H7.2.1.4), where the steel yield stress Fy=58.6 ksi. Residual stresses and
initial plastic strains, as discussed in Section 1675H7.2.1.6, were not considered in the ABAQUS
models because their implementation requires further validation and they were not
observed to markedly influence column ultimate strength (see 1676HFigure 7.48 and 1677HFigure
7.49).
Imperfections were imposed on the initial column geometry in ABAQUS with
custom Matlab code which combines the local, distortional buckling, and global cross‐
section mode shapes from CUFSM along the column length. Two simulations were
performed for each column, one model with 25% CDF local and distortional
imperfection magnitudes and L/2000 global imperfections (where L is the length of the
column) and the other model with 75% CDF local and distortional imperfection
magnitudes and a global imperfection magnitude of L/1000 (see Section 1678H7.2.1.5 for local
and distortional imperfection definitions).
The global imperfection magnitude assumptions are based on hot‐rolled column out‐
of‐straightness measurements (Galambos 1998b) because no formal guidelines are
currently available for cold‐formed steel columns. The use of hot‐rolled steel column
imperfection magnitudes is consistent with the DSM approach for global buckling
317
controlled failures. DSM employs the same global design curve as that specified by the
Structural Stability Research Council (SSRC) for hot‐rolled steel (Galambos 1998b),
thereby indirectly assuming that the influence of hot‐rolled steel global imperfection
magnitudes also apply to cold‐formed steel. The global imperfection shape of the
columns in the simulation database was either weak‐axis flexural buckling or flexural‐
torsional buckling, depending on the cross‐section dimensions and length of the column.
C‐sections are not symmetric about their weak bending axis, and therefore the direction
of the global imperfection influences the predicted strength when weak‐axis flexural
buckling defines the global imperfection shape (e.g., web in compression from bowing
or flange lips in compression from bowing). Simulations with both ± L/1000 and ±
L/2000 imperfection magnitudes were performed to capture this strength effect for
weak‐axis flexural buckling mode shapes. Global imperfections were not considered for
columns with L/D≤18 (i.e., stockier columns with a low sensitivity to global
imperfections), where D is the out‐to‐out flange width of the column.
The local (Pcrl), distortional (Pcrd), and global (Pcre) critical elastic buckling loads were
predicted for each column with custom Matlab code based on the CUFSM prediction
methods described in Section 1679H4.2.7. The database of simulated column experiments,
including cross‐section type, column and hole geometry, simulated ultimate strength
(Ptest25 and Ptest75) and critical elastic buckling loads for each column (including the
presence of holes) is provided in 1680HAppendix K.
318
1.2 104BDistortional buckling study
A group of 20 columns from the SSMA column simulation database was chosen to
evaluate the influence of the ratio Anet/Ag on the tested strength of columns predicted to
collapse with a distortional failure mode. Ag is the gross cross‐sectional area of a column
and Anet is the cross‐sectional area at the location of a hole. In this study the column
length, L, is held constant at 24 in. and the column widths range from 6 in. to 12 in. The
SSMA cross‐sections chosen have relatively thick sheet steel (t up to 0.1017 in.) which
prevents a local buckling type failure. The web of each column has two circular holes
where the hole spacing S=12 in (see 1681HFigure 3.2 for the definition of S). The hole depth
(diameter), hhole, is varied for each column to produce Anet/Ag of 1.0 (no holes), 0.9, 0.8, 0.7,
and 0.6. Refer to 1682HAppendix K, Study Type D, for specific cross‐section and hole
geometry information for each column. 1683HFigure 8.2 provides an example of an SSMA
600S250‐97 structural stud column considered in the study.
1.0 0.90 0.80 0.70 0.60Anet/Ag
SSMA 800S250-97 structural stud column
Figure 8.2 SSMA 800S250‐97 structural stud with web holes considered in the DSM distortional buckling study
The simulation results for Anet/Ag =1.0, 0.9, 0.8, 0.7, and 0.6 are compared to the DSM
distortional buckling prediction curve in 1684HFigure 8.4 to 1685HFigure 8.8. The column strengths,
Ptest25 and Ptest75, without holes (Anet/Ag =1.0) are consistent with the DSM design curve as
319
shown in 1686HFigure 8.4a, confirming the viability of the nonlinear simulation protocol. The
mean and standard deviation of the simulated test to predicted ratio is 1.10 and 0.10
respectively for 25% CDF local and distortional imperfections, and 1.06 and 0.13 for 75%
CDF imperfections (global imperfections are not considered in these stocky columns).
For the columns with holes, the simulated test strengths diverge from the DSM
prediction curve as distortional slenderness, λd=(Pyg/Pcrd)0.5, decreases as shown in 1687HFigure
8.5a to 1688HFigure 8.8a (Pyg is the squash load of the column calculated with the gross cross‐
sectional area Ag). This divergent trend in Ptest with decreasing λd can be explained as
follows. When λd is high (i.e. Pcrd is low relative to Pyg), the column strength is lower than
Pyg because the collapse mechanism is controlled by distortional buckling deformations.
The presence of a hole may decrease Pcrd (as predicted with the method in Section
1689H4.2.7.2), but the distortional failure mechanism still dominates in this case. When λd is
low, Pcrd is much higher than Pyg and the column is not as sensitive to distortional
deformation. Instead, the column fails by yielding of the cross‐section. When a hole is
added, the yielding of the cross‐section occurs at the location of the hole (i.e., at the net
section) resulting in the collapse of the unstiffened strips adjacent to the hole. This
collapse is accompanied by distortional and global deformations caused by the
reduction in stiffness at the net section. These two column failure mechanisms, a
distortional buckling failure (when λd is high) and yielding and collapse of the net
section (when λd is low), are compared in 1690HFigure 8.3.
320
1.0 0.90 0.80 0.70 0.60Anet/Ag
Distortional buckling failureYielding and collapse of the unstiffened strips adjacent to the holes accompanied by distortional and global (weak axis flexural) deformation
0.60 0.59 0.57 0.54 0.45Ptest25/Pyg
1.33 1.36 1.39 1.42 1.45λd
Figure 8.3 SSMA 800S250‐97 structural stud failure mode transition from distortional buckling to yielding at the net section
The observations from this study are used to formulate a modified DSM distortional
curve for columns with holes which captures the failure mechanism transition from
yielding at the net cross‐section to a distortional type failure mode and limits the
strength of the column to its squash load at the net section:
Distortional Buckling The nominal axial strength, Pnd, for distortional buckling shall be calculated in accordance with the following:
(a) For λd ≤λd1
Pnd = ynetP (cap on column strength) (b) For λd1<λd≤λd2
Pnd = d
1d2d
2dynetynet
PPP λ
λλ ⎟⎟⎠
⎞⎜⎜⎝
⎛−
−− (yield control transition)
(c) For λd > λd2
Pnd = y
6.0
y
crd
6.0
y
crd PPP
PP
25.01 ⎟⎟⎠
⎞⎜⎜⎝
⎛
⎟⎟⎟
⎠
⎞
⎜⎜⎜
⎝
⎛
⎟⎟⎠
⎞⎜⎜⎝
⎛− (existing DSM distortional curve)
where λd = crdy PP
λd1 = )PP(561.0 yynet
λd2 = ( )( )13PP14561.0 4.0yynet −−
Pd2 = ( )( )( ) y2.1
2d2.1
2d P1125.01 λλ−
Pynet = FyAnet≥0.6Py
Anet = Column cross-sectional area at the location of hole(s) Pcrd = Critical elastic distortional column buckling load including hole(s)
321
The modified DSM distortional curve is added in 1691HFigure 8.5b to 1692HFigure 8.8b as Anet/Ag
decreases, simulating the transition from the existing DSM curve to the capped column
strength exhibited by the simulated test data. The linear portion of the modified
prediction curve represents the unstiffened strip distortional collapse mechanism and
the nonlinear portion represents a collapse mechanism driven by distortional buckling.
This proposed modification to the DSM distortional prediction curve will be compared
against the column experiments database developed in Section 1693H4.2.6.1 as a part of several
proposed DSM options considered later in this chapter.
0 0.5 1 1.5 2 2.5 3 3.5 40
0.2
0.4
0.6
0.8
1
1.2
1.4
distortional slenderness, λd=(Py/Pcrd)0.5
Pte
st/P
y
DSM (no hole)FE 25% CDF imperfectionsFE 75% CDF imperfections
0 0.5 1 1.5 2 2.5 3 3.5 40
0.2
0.4
0.6
0.8
1
1.2
1.4
distortional slenderness, λd=(Py/Pcrd)0.5
Pte
st/P
y
DSM (no hole)DSM (proposed, with holes)FE 25% CDF imperfectionsFE 75% CDF imperfections
Figure 8.4 Comparison of simulated column strengths (Anet/Ag=1.0) to (a) the existing DSM distortional buckling design curve and to (b) the proposed DSM distortional buckling curve for columns with holes
0 0.5 1 1.5 2 2.5 3 3.5 40
0.2
0.4
0.6
0.8
1
1.2
1.4
distortional slenderness, λd=(Py/Pcrd)0.5
Pte
st/P
y
DSM (no hole)FE 25% CDF imperfectionsFE 75% CDF imperfections
0 0.5 1 1.5 2 2.5 3 3.5 40
0.2
0.4
0.6
0.8
1
1.2
1.4
distortional slenderness, λd=(Py/Pcrd)0.5
Pte
st/P
y
DSM (no hole)DSM (proposed, with holes)FE 25% CDF imperfectionsFE 75% CDF imperfections
Figure 8.5 Comparison of simulated column strengths (Anet/Ag=0.90) to (a) the existing DSM distortional buckling design curve and to (b) the proposed DSM distortional buckling curve for columns with holes
322
0 0.5 1 1.5 2 2.5 3 3.5 40
0.2
0.4
0.6
0.8
1
1.2
1.4
distortional slenderness, λd=(Py/Pcrd)0.5
Pte
st/P
y
DSM (no hole)FE 25% CDF imperfectionsFE 75% CDF imperfections
0 0.5 1 1.5 2 2.5 3 3.5 40
0.2
0.4
0.6
0.8
1
1.2
1.4
distortional slenderness, λd=(Py/Pcrd)0.5
Pte
st/P
y
DSM (no hole)DSM (proposed, with holes)FE 25% CDF imperfectionsFE 75% CDF imperfections
Figure 8.6 Comparison of simulated column strengths (Anet/Ag=0.80) to (a) the existing DSM distortional buckling design curve and to (b) the proposed DSM distortional buckling curve for columns with holes
0 0.5 1 1.5 2 2.5 3 3.5 40
0.2
0.4
0.6
0.8
1
1.2
1.4
distortional slenderness, λd=(Py/Pcrd)0.5
Pte
st/P
y
DSM (no hole)FE 25% CDF imperfectionsFE 75% CDF imperfections
0 0.5 1 1.5 2 2.5 3 3.5 40
0.2
0.4
0.6
0.8
1
1.2
1.4
distortional slenderness, λd=(Py/Pcrd)0.5
Pte
st/P
y
DSM (no hole)DSM (proposed, with holes)FE 25% CDF imperfectionsFE 75% CDF imperfections
Figure 8.7 Comparison of simulated column strengths (Anet/Ag=0.70) to (a) the existing DSM distortional buckling design curve and to (b) the proposed DSM distortional buckling curve for columns with holes
0 0.5 1 1.5 2 2.5 3 3.5 40
0.2
0.4
0.6
0.8
1
1.2
1.4
distortional slenderness, λd=(Py/Pcrd)0.5
Pte
st/P
y
DSM (no hole)FE 25% CDF imperfectionsFE 75% CDF imperfections
0 0.5 1 1.5 2 2.5 3 3.5 40
0.2
0.4
0.6
0.8
1
1.2
1.4
distortional slenderness, λd=(Py/Pcrd)0.5
Pte
st/P
y
DSM (no hole)DSM (proposed, with holes)FE 25% CDF imperfectionsFE 75% CDF imperfections
Figure 8.8 Comparison of simulated column strengths (Anet/Ag=0.60) to (a) the existing DSM distortional buckling design curve and to (b) the proposed DSM distortional buckling curve for columns with holes
323
1.3 105BGlobal buckling study
This study compares simulated strengths to DSM predictions of cold‐formed steel
columns with holes predicted to experience a global failure. A global failure is triggered
by yielding for a stocky column and flexural or flexural‐torsional buckling for slender
columns. No modifications are proposed to the DSM global buckling design curve for
columns with holes, as the influence of holes on short columns will be accounted for
with the DSM local buckling design curve (see Section 1694H8.1.4). For example, when Pne=Pyg,
Pnl will always be made less than or equal to Pynet, and therefore the nominal column
strength, Pn, will always be less than or equal to Pynet.
A group of 18 columns predisposed to a global failure were selected from the SSMA
column simulation database. In this study the column length, L, varied from 8 in. to 96
in. to consider a wide range of global column slenderness, λc=(Pyg/Pcre)0.5. The SSMA
cross‐sections are purposely selected with low local buckling slenderness (i.e., sections
with thicker sheet steel up to t=0.1017 in. and relatively narrow flanges and webs). DSM
predicts that local buckling does not influence global buckling behavior when λl ≤0.776.
The web of each column contains evenly spaced slotted holes where the hole spacing S
varies from 8 in. to 22 in. The hole length, Lhole, is held constant at 4 in., while the hole
depth, hhole, is varied for each column to produce Anet/Ag of 1.0 (no holes), 0.9, and 0.8.
(Refer to 1695HAppendix K, Study Type G, for specific column cross‐section and hole
geometry information.) The four columns with the lowest global slenderness (for
example, Specimen ID # 137 to 140 in 1696HAppendix K) were modeled with circular holes
instead of slotted holes because the slotted holes resulted in impractical column layouts,
324
with the hole extending over more than 50% of the column length. The global
imperfection shape for five of the longer columns was weak‐axis flexural buckling, and
therefore four simulated strengths are determined for these columns (instead of the
typical two): 25% CDF local and distortional imperfections with ±L/2000 global
imperfections and 75% CDF local and distortional imperfections with ±L/1000 global
imperfections.
1697HFigure 8.9 to 1698HFigure 8.11 compare the simulated column strengths to the DSM global
prediction curve as Anet/Ag decreases. The simulated strengths for columns without holes
are consistent with the DSM global prediction curve as shown in 1699HFigure 8.9a. The mean
and standard deviation of the simulated test to predicted ratio for columns without
holes is 1.06 and 0.05 respectively for 25% CDF local and distortional imperfections ±
L/2000 global imperfection and 0.95 and 0.07 for 75% local and distortional
imperfections ± L/1000 global imperfection.
1700HFigure 8.10a and 1701HFigure 8.11a demonstrate that for columns with holes, the predicted
strengths are consistent with the DSM global design curve when global slenderness λc is
greater than 2. Most of the columns in this region fail by weak‐axis flexural buckling.
When λc is between 1 and 2, all of the columns fail by flexural‐torsional buckling and the
simulated column strengths (with 25% CDF imperfections) are 20% higher than the DSM
predictions. This conservative trend is caused by the simplified prediction method
developed in Section 1702H4.3.2.3, which is know to be a conservative predictor of Pcre when
torsional buckling influences the global buckling mode. When Pcre is underpredicted, the
global slenderness increases, which shifts the tested data off of the DSM design curve;
325
the shift is especially clear in 1703HFigure 8.11a. This observation further motivates the future
work to study the influence of holes on the warping torsion constant, Cw.
Simulated column strengths diverge below the DSM prediction curve when λc
decreases and Anet/Ag increases as shown in 1704HFigure 8.10a and 1705HFigure 8.11a. These
columns are short, ranging in length from 8 in. to 26 in., and exhibit a yielding failure
mode at the net section, similar to that observed in the distortional failure study in
1706HFigure 8.3. This observation supports the proposed modification to the DSM
distortional buckling curve, which accurately predicts the strengths of these columns as
shown in 1707HFigure 8.9b and 1708HFigure 8.11b, where the diverging data points are plotted
against the modified DSM distortional prediction curve. This observation reiterates the
conclusion drawn in the distortional buckling study, that yielding and collapse of the
unstiffened strips adjacent to a hole influence both distortional and global failure modes
as slenderness decreases.
0 0.5 1 1.5 2 2.5 3 3.5 40
0.2
0.4
0.6
0.8
1
1.2
1.4
global slenderness, λc=(Py/Pcre)0.5
Pte
st/P
y
DSM (no hole)no holes, FE 25% CDF imperfectionsno holes, FE 75% CDF imperfections
0 0.5 1 1.5 2 2.5 3 3.5 40
0.2
0.4
0.6
0.8
1
1.2
1.4
distortional slenderness, λd=(Py/Pcrd)0.5
Pte
st/P
y
DSM (no hole)no holes, FE 25% CDF imperfectionsno holes, FE 75% CDF imperfections
Figure 8.9 Comparison of simulated column strengths (Anet/Ag=1.00) to (a) the existing DSM global buckling design curve and to (b) the proposed DSM distortional buckling curve for columns with holes
326
0 0.5 1 1.5 2 2.5 3 3.5 40
0.2
0.4
0.6
0.8
1
1.2
1.4
global slenderness, λc=(Py/Pcre)0.5
Pte
st/P
y
DSM (no hole)slotted holes, FE 25% CDF imperfectionsslotted holes, FE 75% CDF imperfectionscircular holes, FE 25% CDF imperfectionscircular holes, FE 75% CDF imperfections
0 0.5 1 1.5 2 2.5 3 3.5 40
0.2
0.4
0.6
0.8
1
1.2
1.4
distortional slenderness, λd=(Py/Pcrd)0.5
Pte
st/P
y
DSM (no hole)slotted holes, FE 25% CDF imperfectionsslotted holes, FE 75% CDF imperfections
Figure 8.10 Comparison of simulated column strengths (Anet/Ag=0.90) to (a) the existing DSM global buckling design curve and to (b) the proposed DSM distortional buckling curve for columns with holes
0 0.5 1 1.5 2 2.5 3 3.5 40
0.2
0.4
0.6
0.8
1
1.2
1.4
global slenderness, λc=(Py/Pcre)0.5
Pte
st/P
y
DSM (no hole)slotted holes, FE 25% CDF imperfectionsslotted holes, FE 75% CDF imperfectionscircular holes, FE 25% CDF imperfectionscircular holes, FE 75% CDF imperfections
0 0.5 1 1.5 2 2.5 3 3.5 40
0.2
0.4
0.6
0.8
1
1.2
1.4
distortional slenderness, λd=(Py/Pcrd)0.5
Pte
st/P
y
DSM (no hole)slotted holes, FE 25% CDF imperfectionsslotted holes, FE 75% CDF imperfections
Figure 8.11 Comparison of simulated column strengths (Anet/Ag=0.80) to (a) the existing DSM global buckling design curve and to (b) the proposed DSM distortional buckling curve for columns with holes
1.4 106BLocal buckling study
The distortional buckling failure study in Section 1709H8.1.2 and the global buckling
failure study in Section 1710H8.1.3 demonstrated that the presence of holes decreases ultimate
strength when cold‐formed steel columns fail by yielding and collapse of the unstiffened
strips adjacent to a hole at the net cross‐section. Holes were observed to have a minimal
influence on ultimate strength when the column failure mode was dictated by elastic
buckling. The goal of this study is to determine if this trend is consistent for columns
with holes experiencing local‐global buckling interaction at failure.
327
Eleven columns from the simulation database in 1711HAppendix K were chosen for this
study. The columns have SSMA cross‐sections and lengths which result in a local
buckling slenderness, λl, ranging from 0.8 to 3.0. The column length, L, varies from 24
in. to 88 in. and column widths range from 3.5 in. to 12 in. The web of each column
contains evenly spaced circular holes where the hole spacing S varies from 12 in. to 17
in. The hole depth (diameter), hhole, is varied for each column to produce Anet/Ag of 1.0 (no
holes), 0.80, and 0.65. Refer to 1712HAppendix K, Study Type L, for specific column cross‐
section and hole geometry information.
The simulated ultimate strengths of the 11 columns without holes, Ptest, are compared
to the DSM local buckling strength prediction, Pnl , in 1713HFigure 8.12. The simulated test to
predicted ratios are more variable than those observed in the distortional and global
failure studies but on average are close to unity, with a trend towards increasingly
conservative predictions with increasing λl as shown in 1714HFigure 8.12a. The mean and
standard deviation of the simulated test to prediction ratio is 1.05 and 0.14 respectively
for 25% CDF local and distortional imperfections ± L/2000 global imperfections and 1.03
and 0.15 for 75% local and distortional imperfections ± L/1000 global imperfections.
328
0 0.5 1 1.5 2 2.5 3 3.5 4
0
0.2
0.4
0.6
0.8
1
1.2
1.4
local slenderness, λl=(Pne/Pcrl)
0.5
Pte
st/P
n l
no holes, FE 25% CDF imperfectionsno holes, FE 75% CDF imperfections
0 0.5 1 1.5 2 2.5 3 3.5 40
0.2
0.4
0.6
0.8
1
1.2
1.4
Pte
st/P
n l
global slenderness, λc=(Py/Pcre)0.5
no holes, FE 25% CDF imperfectionsno holes, FE 75% CDF imperfections
Figure 8.12 Comparison of column test‐to‐prediction ratios for columns (Anet/Ag=1.0) failing by local‐global buckling interaction as a function of (a) local slenderness (b) global slenderness
1715HFigure 8.13 and 1716HFigure 8.14 compare the simulated strengths of the 11 columns to the
predicted strength, Pnl , as Anet/Ag decreases from 1.0 (no hole), to 0.80, to 0.65. (In 1717HFigure
8.13 and 1718HFigure 8.14 the test strengths are those associated with Ptest25+ in 1719HAppendix K, i.e.
the 25% CDF local and distortional imperfection magnitudes and +L/2000 global
imperfection magnitudes.) 1720HFigure 8.13a compares the simulated strengths Ptest25+ to Pnl
without the influence of holes. (The local‐global buckling interaction complicates the
comparison because Pnl and λl are both a function of Pne. By initially assuming that Pne is
not influenced by the hole, the effect of hole size on simulated strength is more clearly
observed.) As local slenderness (λl) decreases in 1721HFigure 8.13a (i.e., the influence of local
buckling on member strength decreases) the tested strength becomes more sensitive to
increasing hole size (i.e., decreasing Anet/Ag), diverging below the prediction Pnl by as
much as 40% when λl=0.75.
1722HFigure 8.14a demonstrates that the sensitivity of column strength to a decrease in Anet/Ag
is related to the ratio of Pynet to Pne. When Pynet/Pne is high, the strength sensitivity to Anet/Ag
329
is low because global buckling initiates the column failure. As Pne/Pynet approaches unity,
column failure is initiated by unstiffened strip buckling and yielding at the net cross‐
section and therefore the sensitivity of column strength to Anet/Ag increases. A column
with the largest drop in strength with increasing hole size is the SSMA 350S162‐68
column with L=34 in. and S=17 in. shown in 1723HFigure 8.15. In this case a large hole
(Anet/Ag=0.65) causes the collapse of the net section resulting in an unfavorable and
sudden weak‐axis flexural failure and a 42% strength reduction when compared to the
same column without holes. The SSMA 350S162‐68 column with smaller holes
(Anet/Ag=0.80) fails in a combination of distortional and flexural‐torsional buckling with a
12% strength reduction.
1724HFigure 8.13b plots the same information as 1725HFigure 8.14a, except now Pne is calculated
using Pcre including the influence of holes. For 8 out of the 11 columns, the prediction
Pnl shifts from unconservative to slightly conservative, even for large holes. One
exception is the SSMA 800S250‐43 column with L=74 in. and S=12 in. shown in 1726HFigure
8.16, where the strength prediction becomes overly conservative as Anet/Ag increases. Pcre
is predicted to decrease by 45% when Anet/Ag=0.65, although the tested strength decreases
by only 10%. 1727HFigure 8.16 demonstrates that the C‐section web is susceptible to local
buckling, and that the presence of holes does not adversely affect the failure mode in
this case. The strength predictions for the SSMA 350S162‐68 column ( 1728HFigure 8.15) and
the SSMA 350S162‐54 column with L=24 in. and S=12 in. are viable when Anet/Ag=0.80, but
are underestimated by 20% with Option 4 (5) when Anet/Ag=0.65 because of the
introduction of an unstable weak‐axis flexural failure mode triggered by the collapse of
330
the net section. A hinge forms at the location of the net section, and the global
slenderness is high enough that the column becomes susceptible to a flexural buckling
mode. These “hinge” failures are not observed in the distortional buckling study (see
Section 1729H8.1.2) because the global slenderness of the columns is lower (i.e. the weak axis
flexural stiffness is higher), avoiding a global buckling failure. Option 6 accurately
predicts the strength of the SSMA 350S162‐68 column and the SSMA 350S162‐54
columns because the method assumes that the global strength, Pne, is reduced by Pynet/Py.
0 0.5 1 1.5 2 2.5 3 3.5 40
0.5
1
1.5
Pte
st/P
n l
local slenderness, λl=(Pne/Pcrl
)0.5
0.65
Anet/Ag
0.80
1.0
0 0.5 1 1.5 2 2.5 3 3.5 40
0.5
1
1.5P
test
/Pn l
local slenderness, λl=(Pne/Pcrl)
0.5
SSMA 350S162-68 column
SSMA 800S250-43 column
Figure 8.13 Comparison of column test‐to‐prediction ratios for columns failing by local‐global buckling interaction with Pne calculated (a) without the influence of holes (b) and with the influence of holes
331
0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 50
0.5
1
1.5
Pte
st/P
n l
Pynet/Pne
Increase in point size, Anet/Ag=1.0,0.80,0.65
0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 50
0.5
1
1.5
Pte
st/P
n l
Pynet/Pne
Figure 8.14 Comparison of column test‐to‐prediction ratios for columns failing by local‐global buckling interaction as a function of Pynet/Pne where Pne is calculated (a) without the influence of holes (b) and with the
influence of holes
1.0 0.80 0.65Anet/Ag
1.0 0.88 0.58Ptest/Ptest,no hole
Distortional, flexural-torsional failure
Distortional, flexural-torsional failure Weak-axis flexural failure
Figure 8.15 SSMA 350S162‐68 column failure mode changes from distortional‐flexural torsional buckling failure to weak axis flexure as hole size increases (L=34 in.)
332
1.0 0.80 0.65Anet/Ag
1.0 0.95 0.90Ptest/Ptest,no hole
Local – global interaction at failure
Unstiffened strip buckling –global interaction at failure
Local – global interaction at failure
Figure 8.16 SSMA 800S250‐43 (L=74 in.) column web local buckling changes to unstiffened strip buckling at peak load as hole size increases
The observations from this study are now employed to propose two options for the
DSM local buckling design curve for columns with holes. The presence of holes
influenced the tested strength of the cold‐formed steel columns over the full range of
local slenderness considered. This result was different from the distortional and global
failure studies, where holes were observed to reduce strength only from the collapse and
yielding at the net section as slenderness decreased. The strength reduction from the
holes was predicted in DSM for 8 out of the 11 columns, when Pnl was calculated with
Pne included the influence of holes (compare 1730HFigure 8.13a to 1731HFigure 8.13b). A transition
similar to that proposed for the DSM distortional design curve is still justified though,
especially when Pynet/Pne≤1 (see 1732HFigure 8.14b), to capture the yielding and collapse at the
net section observed in columns with low local and global slenderness. The strength of
two columns with large holes were underpredicted because of unstable global collapse
initiated by yielding at the net‐section, motivating the implementation of a limit on hole
333
size (i.e., Anet/Ag) to ensure the viability of the DSM approach. Two modification options
are proposed for the DSM local design curve based on these conclusions:
Local Buckling (Option A) The nominal axial strength, Pnl, for local buckling shall be calculated in accordance with the following:
(a) For λl ≤λl1 Pnl = Pne≤ Pynet (cap on column strength)
(b) For λl1<λl≤λl2
Pnl = ( )112
2ynetynet
PPP ll
ll
l λλλλ
−⎟⎟⎠
⎞⎜⎜⎝
⎛−
−− (yield transition when Pynet/Pne ≤1)
(c) For λl > λl2
Pnl = ne
4.0
ne
cr4.0
ne
cr PPP
PP
15.01 ⎟⎟⎠
⎞⎜⎜⎝
⎛⎟⎟
⎠
⎞
⎜⎜
⎝
⎛⎟⎟⎠
⎞⎜⎜⎝
⎛− ll (DSM local buckling curve, unchanged)
where λl =
lcrne PP
λl1 = )PP(776.0 neynet≤0.776
λl2 = ( )( )7.0PP7.1776.0 6.1neynet −− , Pynet/Pne≤1
= 0.776, Pynet/Pne>1 (no transition when Pynet/Pne >1 ) Pl2 = ( )( )( ) ne
8.02
8.02 P1115.01 ll λλ−
Pynet = FyAnet≥0.6Py (limit reduction of the net section to 0.6Py)
Anet = Column cross-sectional area at the location of hole(s) Pcrl = Critical elastic local column buckling load including hole(s)
Local Buckling (Option B) The nominal axial strength, Pnl, for local buckling shall be calculated in accordance with the following:
(a) For λl ≤λl1 Pnl = Pynet (Pne /Py ) (cap on column strength)
(b) For λl1<λl≤λl2
Pnl = ( ) ( )112
2yneynet
y
neynet
PPPPPPP ll
ll
l λλλλ
−⎟⎟⎠
⎞⎜⎜⎝
⎛−
−−⎟
⎟⎠
⎞⎜⎜⎝
⎛ (yield transition when Pynet/Pne ≤1)
(c) For λl > λl2
Pnl = ne
4.0
ne
cr4.0
ne
cr PPP
PP
15.01 ⎟⎟⎠
⎞⎜⎜⎝
⎛⎟⎟
⎠
⎞
⎜⎜
⎝
⎛⎟⎟⎠
⎞⎜⎜⎝
⎛− ll (DSM local buckling curve, unchanged)
where λl =
lcrne PP
λl1 = )PP(776.0 yynet
λl2 = ( )( )7.0PP7.1776.0 6.1yynet −−
Pl2 = ( )( )( ) ne8.0
28.0
2 P1115.01 ll λλ− Pynet = FyAnet≥0.6Py (limit reduction of the net section to 0.6Py)
334
Anet = Column cross-sectional area at the location of hole(s) Pcrl = Critical elastic local column buckling load including hole(s)
Option A imposes a transition from the DSM local buckling curve to column
strength at the net section, Pynet, when Pynet< Pne as shown in 1733HFigure 8.17a for the case when
Pne=Py (i.e., stub columns) and Pynet=0.8Pyg. When Pynet>Pne, Option A assumes that holes
influence only the critical elastic buckling loads (Pcrl, Pcre) and otherwise do not change
the failure mode of the column; this case is demonstrated in 1734HFigure 8.17c when Pcre = Pyg.
Option B also imposes a transition to the net column strength from the DSM local failure
curve, although the transition is assumed to occur for all values of Pynet/Pne. In essence,
the Option B curve for stub columns shown in 1735HFigure 8.17a is scaled down based on the
ratio Pynet/Py. The result is an additional reduction in predicted strength for global
column failures without local buckling interaction that is not captured by Option A.
This difference between Option A and Option B is highlighted in 1736HFigure 8.17b, where
Pynet=0.8Pyg and Pcre = 5Pyg. The validity of both options are evaluated in the following
section against the simulation database and the experiment database assembled in
1737HChapter 4.
0 0.5 1 1.5 2 2.5 3 3.5 40
0.2
0.4
0.6
0.8
1
1.2
1.4
local slenderness, λl=(Pne/Pcrl
)0.5
Pn l
/Py
DSM local curve (no hole)DSM local curve (Option A)DSM local curve (Option B)
0 0.5 1 1.5 2 2.5 3 3.5 40
0.2
0.4
0.6
0.8
1
1.2
1.4
local slenderness, λl=(Pne/Pcrl
)0.5
Pn l
/Py
DSM local curve (no hole)DSM local curve (Option A)DSM local curve (Option B)
0 0.5 1 1.5 2 2.5 3 3.5 40
0.2
0.4
0.6
0.8
1
1.2
1.4
local slenderness, λl=(Pne/Pcrl
)0.5
Pn l
/Py
DSM local curve (no hole)DSM local curve (Option A)DSM local curve (Option B)
Pcre=100Pyg
Pynet=0.8Pyg
Pcre=5Pyg
Pynet=0.8Pyg
Pcre=Pyg
Pynet=0.8Pyg
Figure 8.17 Comparison of DSM local buckling design curve options when Pynet=0.80 Pyg and (a) Pcre=100Pyg, (b) Pcre=5Pyg, and (c) Pcre=Pyg
335
1.5 107BPresentation and evaluation of DSM options
Six options for extending DSM to columns with holes are evaluated in this section.
The options range from simple substitutions in the existing code to more involved
modifications, including the incorporation of the design curve transitions discussed in
Section 1738H8.1.2 and Section 1739H8.1.4 for distortional and local buckling.
1.5.1 182BDescription of DSM options
Option 1: Include hole(s) in Pcr determinations, ignore hole otherwise This method, in presentation, appears identical to currently available DSM expressions
Flexural, Torsional, or Torsional-Flexural Buckling The nominal axial strength, Pne, for flexural, … or torsional- flexural buckling is
for 5.1c ≤λ Pne = yP658.02c ⎟
⎠⎞⎜
⎝⎛ λ
for λc > 1.5 crey2c
ne P877.0P877.0P =⎟⎟⎠
⎞⎜⎜⎝
⎛
λ=
where λc = crey PP
Py = AgFy Pcre= Critical elastic global column buckling load … (including hole(s)) Ag = gross area of the column
Local Buckling The nominal axial strength, Pnl, for local buckling is
for λl 776.0≤ Pnl = Pne
336
for λl > 0.776 Pnl = ne
4.0
ne
cr4.0
ne
cr PPP
PP
15.01 ⎟⎟⎠
⎞⎜⎜⎝
⎛⎥⎥
⎦
⎤
⎢⎢
⎣
⎡⎟⎟⎠
⎞⎜⎜⎝
⎛− ll
where λl = lcrne PP Pcrl = Critical elastic local column buckling load … (including hole(s)) Pne is defined above.
Distortional Buckling The nominal axial strength, Pnd, for distortional buckling is
for λd 561.0≤ Pnd = Py
for λd > 0.561 Pnd = y
6.0
y
crd
6.0
y
crd PPP
PP
25.01 ⎟⎟⎠
⎞⎜⎜⎝
⎛
⎟⎟⎟
⎠
⎞
⎜⎜⎜
⎝
⎛
⎟⎟⎠
⎞⎜⎜⎝
⎛−
where λd = crdy PP
Pcrd = Critical elastic distortional column buckling load … (including hole(s))
337
Option 2: Include hole(s) in Pcr determinations, Use Pynet everywhere The only change in this method is to replace Py with Pynet
Flexural, Torsional, or Torsional-Flexural Buckling The nominal axial strength, Pne, for flexural, … or torsional- flexural buckling is
for 5.1cnet ≤λ Pne = ( ) ynetPcnet2
658.0 λ
for λcnet > 1.5 creynet
cnetne PPP 877.0877.0
2=⎟⎟
⎠
⎞⎜⎜⎝
⎛=
λ
where λcnet = creynet PP
Pynet = AnetFy Pcre= Critical elastic global column buckling load … (including hole(s)) Anet = net area of the column
Local Buckling The nominal axial strength, Pnl, for local buckling is
for λl 776.0≤ Pnl = Pne
for λl > 0.776 Pnl = ne
4.0
ne
cr4.0
ne
cr PPP
PP
15.01 ⎟⎟⎠
⎞⎜⎜⎝
⎛⎥⎥
⎦
⎤
⎢⎢
⎣
⎡⎟⎟⎠
⎞⎜⎜⎝
⎛− ll
where λl = lcrne PP Pcrl = Critical elastic local column buckling load … (including hole(s)) Pne is defined above
Distortional Buckling The nominal axial strength, Pnd, for distortional buckling is
for λdnet 561.0≤ Pnd = Pynet
for λdnet > 0.561 Pnd = ynet
6.0
ynet
crd6.0
ynet
crd PPP
PP
25.01 ⎟⎟⎠
⎞⎜⎜⎝
⎛
⎟⎟⎟
⎠
⎞
⎜⎜⎜
⎝
⎛
⎟⎟⎠
⎞⎜⎜⎝
⎛−
where λdnet = crdynet PP
Pcrd = Critical elastic distortional column buckling load … (including hole(s))
338
Option 3: Cap Pnl and Pnd, otherwise no strength change, include hole(s) in Pcr This method puts bounds in place and assumes local-global interaction happens at full Pne
Flexural, Torsional, or Torsional-Flexural Buckling The nominal axial strength, Pne, for flexural, … or torsional- flexural buckling is
for 5.1c ≤λ Pne = yP658.02c ⎟
⎠⎞⎜
⎝⎛ λ
for λc > 1.5 crey2c
ne P877.0P877.0P =⎟⎟⎠
⎞⎜⎜⎝
⎛
λ=
where λc = crey PP
Pcre= Critical elastic global column buckling load … (including hole(s)) Py = AgFy Ag = gross area of the column
Local Buckling The nominal axial strength, Pnl, for local buckling is
for λl 776.0≤ Pnl = ynetne PP ≤
for λl > 0.776 Pnl = ynetne
4.0
ne
cr4.0
ne
cr PPPP
PP
15.01 ≤⎟⎟⎠
⎞⎜⎜⎝
⎛⎥⎥
⎦
⎤
⎢⎢
⎣
⎡
⎟⎟⎠
⎞⎜⎜⎝
⎛− ll
where λl = lcrne PP Pcrl = Critical elastic local column buckling load … (including hole(s)) Pne is defined in Section above. Pynet = AnetFy Anet = net area of the column
Distortional Buckling The nominal axial strength, Pnd, for distortional buckling is
for λd 561.0≤ Pnd = Py ≤ Pynet
for λd > 0.561 Pnd = ynet.y
6.0
y
crd6.0
y
crd PPP
PP
P25.01 ≤⎟
⎟⎠
⎞⎜⎜⎝
⎛
⎟⎟⎟
⎠
⎞
⎜⎜⎜
⎝
⎛
⎟⎟⎠
⎞⎜⎜⎝
⎛−
where λd = crdy PP
Pcrd = Critical elastic distortional column buckling load … (including hole(s))
339
Option 4: Cap Pnl, transition Pnd, include hole(s) in Pcr determinations This method puts bounds and transition in place, assumes local-global interaction at full Pne
Flexural, Torsional, or Torsional-Flexural Buckling The nominal axial strength, Pne, for flexural, … or torsional- flexural buckling is
for 5.1c ≤λ Pne = yP658.02c ⎟
⎠⎞⎜
⎝⎛ λ
for λc > 1.5 crey2c
ne P877.0P877.0P =⎟⎟⎠
⎞⎜⎜⎝
⎛
λ=
where λc = crey PP
Pcre= Critical elastic global column buckling load … (including hole(s)) Py = AgFy Ag = gross area of the column
Local Buckling The nominal axial strength, Pnl, for local buckling is
for λl 776.0≤ Pnl = ynetne PP ≤
for λl > 0.776 Pnl = ynetne
4.0
ne
cr4.0
ne
cr PPPP
PP
15.01 ≤⎟⎟⎠
⎞⎜⎜⎝
⎛⎥⎥
⎦
⎤
⎢⎢
⎣
⎡
⎟⎟⎠
⎞⎜⎜⎝
⎛− ll
where λl = lcrne PP Pcrl = Critical elastic local column buckling load … (including hole(s)) Pne is defined in Section above. Pynet = AnetFy Anet = net area of the column
Distortional Buckling The nominal axial strength, Pnd, for distortional buckling shall be calculated in accordance with the following:
(a) For λd ≤λd1 Pnd = ynetP (b) For λd1<λd≤λd2
Pnd = d
1d2d
2dynetynet
PPP λ
λλ ⎟⎟⎠
⎞⎜⎜⎝
⎛−
−−
(c) For λd > λd2
Pnd = y
6.0
y
crd
6.0
y
crd PPP
PP
25.01 ⎟⎟⎠
⎞⎜⎜⎝
⎛
⎟⎟⎟
⎠
⎞
⎜⎜⎜
⎝
⎛
⎟⎟⎠
⎞⎜⎜⎝
⎛−
where λd = crdy PP
λd1 = )PP(561.0 yynet
λd2 = ( )( )13PP14561.0 4.0yynet −−
Pd2 = ( )( )( ) y2.1
2d2.1
2d P1125.01 λλ−
Pynet = FyAnet≥0.6Py
Anet = Column cross-sectional area at the location of hole(s) Pcrd = Critical elastic distortional column buckling load including hole(s)
340
Option 5: Transition Pnl (Option A), transition Pnd, include hole(s) in Pcr determinations This method puts bounds and transition in place, assumes local-global interaction at full Pne
Flexural, Torsional, or Torsional-Flexural Buckling The nominal axial strength, Pne, for flexural, … or torsional- flexural buckling is
for 5.1c ≤λ Pne = yP658.02c ⎟
⎠⎞⎜
⎝⎛ λ
for λc > 1.5 crey2c
ne P877.0P877.0P =⎟⎟⎠
⎞⎜⎜⎝
⎛
λ=
where λc = crey PP
Pcre= Critical elastic global column buckling load … (including hole(s)) Py = AgFy Ag = gross area of the column
Local Buckling The nominal axial strength, Pnl, for local buckling shall be calculated in accordance with the following:
(a) For λl ≤λl1 Pnl = Pne≤ Pynet (cap on column strength)
(b) For λl1<λl≤λl2
Pnl = ( )112
2ynetynet
PPP ll
ll
l λλλλ
−⎟⎟⎠
⎞⎜⎜⎝
⎛−
−− (yield transition when Pynet/Pne ≤1)
(c) For λl > λl2
Pnl = ne
4.0
ne
cr4.0
ne
cr PPP
PP
15.01 ⎟⎟⎠
⎞⎜⎜⎝
⎛⎟⎟
⎠
⎞
⎜⎜
⎝
⎛⎟⎟⎠
⎞⎜⎜⎝
⎛− ll (DSM local buckling curve, unchanged)
where λl =
lcrne PP
λl1 = )PP(776.0 neynet≤0.776
λl2 = ( )( )7.0PP7.1776.0 6.1neynet −− , Pynet/Pne≤1
= 0.776, Pynet/Pne>1 (no transition when Pynet/Pne >1 ) Pl2 = ( )( )( ) ne
8.02
8.02 P1115.01 ll λλ−
Pynet = FyAnet≥0.6Py (limit reduction of the net section to 0.6Py)
Anet = Column cross-sectional area at the location of hole(s) Pcrl = Critical elastic local column buckling load including hole(s)
Distortional Buckling
Same as Option 4
341
Option 6: Transition Pnl (Option B), transition Pnd, include hole(s) in Pcr determinations This method puts bounds and transition in place, assumes local-global interaction at full Pne
Flexural, Torsional, or Torsional-Flexural Buckling The nominal axial strength, Pne, for flexural, … or torsional- flexural buckling is
for 5.1c ≤λ Pne = yP658.02c ⎟
⎠⎞⎜
⎝⎛ λ
for λc > 1.5 crey2c
ne P877.0P877.0P =⎟⎟⎠
⎞⎜⎜⎝
⎛
λ=
where λc = crey PP
Pcre= Critical elastic global column buckling load … (including hole(s)) Py = AgFy Ag = gross area of the column
Local Buckling The nominal axial strength, Pnl, for local buckling shall be calculated in accordance with the following:
(a) For λl ≤λl1 Pnl = Pynet (Pne /Py ) (cap on column strength)
(b) For λl1<λl≤λl2
Pnl = ( ) ( )112
2yneynet
y
neynet
PPPPPPP ll
ll
l λλλλ
−⎟⎟⎠
⎞⎜⎜⎝
⎛−
−−⎟
⎟⎠
⎞⎜⎜⎝
⎛ (yield transition when Pynet/Pne ≤1)
(c) For λl > λl2
Pnl = ne
4.0
ne
cr4.0
ne
cr PPP
PP
15.01 ⎟⎟⎠
⎞⎜⎜⎝
⎛⎟⎟
⎠
⎞
⎜⎜
⎝
⎛⎟⎟⎠
⎞⎜⎜⎝
⎛− ll (DSM local buckling curve, unchanged)
where λl =
lcrne PP
λl1 = )PP(776.0 yynet
λl2 = ( )( )7.0PP7.1776.0 6.1yynet −−
Pl2 = ( )( )( ) ne8.0
28.0
2 P1115.01 ll λλ− Pynet = FyAnet≥0.6Py (limit reduction of the net section to 0.6Py)
Anet = Column cross-sectional area at the location of hole(s) Pcrl = Critical elastic local column buckling load including hole(s)
Distortional Buckling
Same as Option 4
342
1.6 108B DSM comparison to column test simulation database
The six DSM prediction options for cold‐formed steel columns with holes are
evaluated with the simulated column experiment database developed in Section 1740H8.1.1
and summarized in 1741HAppendix K. (Tested strengths with and without global
imperfections are provided in 1742HAppendix K. The simulated strengths considered in this
study contain global imperfections, except for stocky columns with L/D<18 where D is
the column flange width). The simulated data is compared against DSM predictions
while evaluating data trends against member slenderness, hole size (Anet/Ag), and column
dimensions L/h, where h is the flat web width of a column.
1743HFigure 8.18 to 1744HFigure 8.21 compare the simulated test data to predictions for local,
distortional, and global buckling controlled column failures. Option 1 is identical to the
existing DSM approach for columns without holes, except the critical elastic buckling
loads (Pcrl, Pcrd, and Pcre) are determined with the influence of holes. Option 1 is observed
to be an accurate predictor of strength when λl, λd, and λc are high, but results in
unconservative predictions (by as much as 30 % for distortional buckling controlled
specimens, see 1745HFigure 8.20) as λl, λd, and λc decreases below 1.5. The unconservative
predictions occur because Option 1 does not account for the column strength limit Pynet,
nor does it account for a transition from an elastic buckling controlled failure to a yield‐
controlled failure at the net section discussed in Section 1746H8.1.2 and Section 1747H8.1.3.
Option 2 is observed to be a conservative predictor in 1748HFigure 8.18 to 1749HFigure 8.21 for
high λl, λd, and λc and demonstrates improved accuracy over Option 1 when slenderness
decreases and hole size increases (see 1750HFigure 8.20). Option 2 replaces Pynet everywhere
343
within the existing DSM formulation, which has the effect of increasing λl, λd, and λc and
decreasing predicted strength. Option 3 test‐to‐predicted trends are similar to Option 1
with increasingly unconservative predictions as slenderness decreases, demonstrating
that the Pynet limit on Pnl and Pnd in Option 3 are not fully effective at capturing the yield
transition to the net section. Option 4 is identical to Option 3 except the yield transition
on the DSM distortional curve developed in Section 1751H8.1.2 is employed to provide a more
accurate prediction of the net‐section yielding influence. Option 4 demonstrates an
improvement in accuracy over Option 3, although it overpredicts the strength of the two
columns discussed in Section 1752H8.1.4 (SSMA 350S162‐68 and SSMA 350S162‐54 columns),
where large holes caused a sudden weak‐axis flexural buckling failure. Option 5
includes both local and distortional yield transitions, although the predictions are
identical to Option 4 because the distortional transition always predicts lower strengths
than the local transition for the columns considered. Option 6 deviates from the other
approaches and accounts for the presence of holes by reducing Pnl by the ratio Pynet/Py
when λl is less than 0.776; this option also always including a local buckling transition
(Option 5 imposes a transition on the DSM local buckling design curve only when
Pynet<Pne, see 1753HFigure 8.17). The reduction in Pnl shifts the global buckling‐controlled
specimens in Options 1 through 5 to the DSM local buckling curve in Option 6, resulting
in conservative predictions with decreasing λl.
1754HTable 8.1 summarizes the test‐to‐predicted ratio statistics for the six DSM options.
The standard deviation (SD) is useful when comparing the methods, because it provides
a metric for how well the trends in strength are following the prediction curves. (The
344
mean is also an important statistic but can hide unconservative prediction trends in
some columns with overconservative predictions in other columns). A low standard
deviation is appealing because it enables higher strength reduction factors in a design
code. The strength reduction factor φ is also provided for each option. φ is calculated
with the following equation from Chapter F of the Specification (AISI‐S100 2007):
( )2222
QPPFMo VVCVVmmm ePFMC +++−= β
φφ , (8.1)
where the calibration coefficient Cφ =1.52 for LRFD, the mean value of the material factor
Mm=1.10 for concentrically loaded compression members, the mean value of the
fabrication factor Fm=1.0, the mean value of the professional factor Pm=1.0, the coefficient
of variation (COV) of the material factor Vm=0.10, the COV of the fabrication factor
Vf=0.05, the COV of the load effect Vq=0.21 for LRFD, and the correction factor Cp=1. The
COV of the test results, Vp, is calculated as the ratio of the standard deviation to the
mean of the test‐to‐predicted statistics in 1755HTable 8.1.
No one option stands out above the rest when studying the table, although Option 2,
3, and 4 (5) have the most evenly distributed statistics between local and distortional
bucking column groups. The observations from this comparison will be combined with
the DSM comparison to the experimental database in the next section.
Table 8.1 DSM test‐to‐predicted statistics for column simulations
Option DescriptionMean SD φ # of tests Mean SD φ # of tests Mean SD φ # of tests
1 Py everywhere 1.06 0.15 0.83 93 1.07 0.17 0.82 178 1.11 0.21 0.78 1142 Pynet everywhere 1.14 0.13 0.86 93 1.24 0.18 0.83 176 1.15 0.18 0.82 1163 Cap Pnl, Pnd 1.06 0.15 0.83 93 1.09 0.17 0.82 186 1.13 0.21 0.79 1064 Transition Pnd, Cap Pnl 1.08 0.14 0.85 89 1.04 0.19 0.79 200 1.16 0.19 0.82 965 Transition Pnd and Pnl (Option A) 1.08 0.14 0.85 89 1.04 0.19 0.79 200 1.16 0.19 0.82 966 Transition Pnd and Pnl (Option B) 1.07 0.20 0.78 221 1.10 0.15 0.85 164 --- --- --- 0
Local buckling Distortional buckling Global buckling
345
0 0.5 1 1.5 2 2.5 30
0.5
1
1.5Option 1 - Py everywhere
λl=(Pne/Pcrl
)0.5
Pn/P
ne
0 0.5 1 1.5 2 2.5 30
0.5
1
1.5Option 2 - Pynet everywhere
λl=(Pne/Pcrl
)0.5
Pn/P
ne
0 0.5 1 1.5 2 2.5 30
0.5
1
1.5Option 3 - cap Pnl
, Pnd
λl=(Pne/Pcrl
)0.5
Pn/P
ne
0 0.5 1 1.5 2 2.5 30
0.5
1
1.5Option 4 - cap Pnl
, transition Pnd
λl=(Pne/Pcrl
)0.5
Pn/P
ne
DSM Local PredictionLocal Controlled Test
0 0.5 1 1.5 2 2.5 30
0.5
1
1.5Option 5 - transition Pnl
(Option A), transition Pnd
λl=(Pne/Pcrl
)0.5
Pn/P
ne
0 0.5 1 1.5 2 2.5 30
0.5
1
1.5Option 6 - transition Pnl
(Option B), transition Pnd
λl=(Pne/Pcrl
)0.5
Pn/P
ne
Figure 8.18 Comparison of simulated test strengths to predictions for columns with local buckling‐controlled failures as a function of local slenderness (tested
strength is normalized by Pne)
346
0 0.5 1 1.5 2 2.5 30
0.5
1
1.5Option 1 - Py everywhere
λl=(Pne/Pcrl
)0.5
Pn/P
y
0 0.5 1 1.5 2 2.5 30
0.5
1
1.5Option 2 - Pynet everywhere
λl=(Pne/Pcrl
)0.5
Pn/P
y
0 0.5 1 1.5 2 2.5 30
0.5
1
1.5Option 3 - cap Pnl
, Pnd
λl=(Pne/Pcrl
)0.5
Pn/P
y
0 0.5 1 1.5 2 2.5 30
0.5
1
1.5Option 4 - cap Pnl
, transition Pnd
λl=(Pne/Pcrl
)0.5
Pn/P
y
DSM Local PredictionLocal Controlled Test
0 0.5 1 1.5 2 2.5 30
0.5
1
1.5Option 5 - transition Pnl
(Option A), transition Pnd
λl=(Pne/Pcrl
)0.5
Pn/P
y
0 0.5 1 1.5 2 2.5 30
0.5
1
1.5Option 6 - transition Pnl
(Option B), transition Pnd
λl=(Pne/Pcrl
)0.5
Pn/P
y
Figure 8.19 Comparison of simulated test strengths to predictions for columns with local buckling‐controlled failures as a function of local slenderness (tested
strength is normalized by Pyg)
347
0 0.5 1 1.5 2 2.5 30
0.5
1
1.5Option 1 - Py everywhere
λd=(Py/Pcrd)0.5
Pn/P
y
0 0.5 1 1.5 2 2.5 30
0.5
1
1.5Option 2 - Pynet everywhere
λdnet=(Pynet/Pcrd)0.5
Pn/P
y
DSM Dist. PredictionDist. Controlled Test
0 0.5 1 1.5 2 2.5 30
0.5
1
1.5Option 3 - cap Pnl
, Pnd
λd=(Py/Pcrd)0.5
Pn/P
y
0 0.5 1 1.5 2 2.5 30
0.5
1
1.5Option 4 - cap Pnl
, transition Pnd
λd=(Py/Pcrd)0.5
Pn/P
y
0 0.5 1 1.5 2 2.5 30
0.5
1
1.5Option 5 - transition Pnl
(Option A), transition Pnd
λd=(Py/Pcrd)0.5
Pn/P
y
0 0.5 1 1.5 2 2.5 30
0.5
1
1.5Option 6 - transition Pnl
(Option B), transition Pnd
λd=(Py/Pcrd)0.5
Pn/P
y
Figure 8.20 Comparison of simulated test strengths to predictions for columns with distortional buckling‐controlled failures as a function of distortional
slenderness
348
0 0.5 1 1.5 2 2.5 30
0.5
1
1.5
λc=(Py/Pcre)0.5
Pn/P
y
Option 1 - Py everywhere
0 0.5 1 1.5 2 2.5 30
0.5
1
1.5
λcnet=(Pynet/Pcre)0.5
Pn/P
y
Option 2 - Pynet everywhere DSM Global Prediction
Global Controlled Test
0 0.5 1 1.5 2 2.5 30
0.5
1
1.5
λc=(Py/Pcre)0.5
Pn/P
y
Option 3 - cap Pnl, Pnd
0 0.5 1 1.5 2 2.5 30
0.5
1
1.5
λc=(Py/Pcre)0.5
Pn/P
y
Option 4 - cap Pnl, transition Pnd
0 0.5 1 1.5 2 2.5 30
0.5
1
1.5
λc=(Py/Pcre)0.5
Pn/P
y
Option 5 - transition Pnl (Option A), transition Pnd
0 0.5 1 1.5 2 2.5 30
0.5
1
1.5
λc=(Py/Pcre)0.5
Pn/P
y
Option 6 - transition Pnl (Option B), transition Pnd
Figure 8.21 Comparison of simulated test strengths to predictions for columns with global buckling‐controlled failures (i.e., no local interaction) as a function of
global slenderness
349
0 0.5 1 1.5 2 2.5 30
0.5
1
1.5
Option 1 - Py everywhere
λl=(Pne/Pcrl
)0.5
Pte
st/P
n
0 0.5 1 1.5 2 2.5 30
0.5
1
1.5
Option 2 - Pynet everywhere
λl=(Pne/Pcrl
)0.5
Pte
st/P
n
Local Controlled
0 0.5 1 1.5 2 2.5 30
0.5
1
1.5
Option 3 - cap Pnl, Pnd
λl=(Pne/Pcrl
)0.5
Pte
st/P
n
0 0.5 1 1.5 2 2.5 30
0.5
1
1.5
Option 4 - cap Pnl, transition Pnd
λl=(Pne/Pcrl
)0.5
Pte
st/P
n
0 0.5 1 1.5 2 2.5 30
0.5
1
1.5
Option 5 - transition Pnl (Option A), transition Pnd
λl=(Pne/Pcrl
)0.5
Pte
st/P
n
0 0.5 1 1.5 2 2.5 30
0.5
1
1.5
Option 6 - transition Pnl (Option B), transition Pnd
λl=(Pne/Pcrl
)0.5
Pte
st/P
n
Figure 8.22 Test‐to‐predicted ratios for local buckling‐controlled simulated column failures as a function of local slenderness
350
0 0.5 1 1.5 2 2.5 30
0.5
1
1.5
Option 1 - Py everywhere
λd=(Py/Pcrd)0.5
Pte
st/P
n
0 0.5 1 1.5 2 2.5 30
0.5
1
1.5
Option 2 - Pynet everywhere
λdnet=(Pynet/Pcrd)0.5
Pte
st/P
n
Distortional Controlled
0 0.5 1 1.5 2 2.5 30
0.5
1
1.5
Option 3 - cap Pnl, Pnd
λd=(Py/Pcrd)0.5
Pte
st/P
n
0 0.5 1 1.5 2 2.5 30
0.5
1
1.5
Option 4 - cap Pnl, transition Pnd
λd=(Py/Pcrd)0.5
Pte
st/P
n
0 0.5 1 1.5 2 2.5 30
0.5
1
1.5
Option 5 - transition Pnl (Option A), transition Pnd
λd=(Py/Pcrd)0.5
Pte
st/P
n
0 0.5 1 1.5 2 2.5 30
0.5
1
1.5
Option 6 - transition Pnl (Option B), transition Pnd
λd=(Py/Pcrd)0.5
Pte
st/P
n
Figure 8.23 Test‐to‐predicted ratios for distortional buckling‐controlled simulated column failures as a function of distortional slenderness
351
0 0.5 1 1.5 2 2.5 30
0.5
1
1.5
Option 1 - Py everywhere
λc=(Py/Pcre)0.5
Pte
st/P
ne
0 0.5 1 1.5 2 2.5 30
0.5
1
1.5
Option 2 - Pynet everywhere
λcnet=(Pynet/Pcre)0.5
Pte
st/P
ne
Global Controlled
0 0.5 1 1.5 2 2.5 30
0.5
1
1.5
Option 3 - cap Pnl, Pnd
λc=(Py/Pcre)0.5
Pte
st/P
ne
0 0.5 1 1.5 2 2.5 30
0.5
1
1.5
Option 4 - cap Pnl, transition Pnd
λc=(Py/Pcre)0.5
Pte
st/P
ne
0 0.5 1 1.5 2 2.5 30
0.5
1
1.5
Option 5 - transition Pnl (Option A), transition Pnd
λc=(Py/Pcre)0.5
Pte
st/P
ne
0 0.5 1 1.5 2 2.5 30
0.5
1
1.5
Option 6 - transition Pnl (Option B), transition Pnd
λc=(Py/Pcre)0.5
Pte
st/P
ne
Figure 8.24 Test‐to‐predicted ratios for simulated global buckling‐controlled column failures (i.e., no local buckling interaction) as a function of global slenderness
352
0.5 0.6 0.7 0.8 0.9 10
0.5
1
1.5
Option 1 - Py everywhere
Anet/Ag
Pte
st/P
n
0.5 0.6 0.7 0.8 0.9 10
0.5
1
1.5
Option 2 - Pynet everywhere
Anet/Ag
Pte
st/P
n
Local Controlled
0.5 0.6 0.7 0.8 0.9 10
0.5
1
1.5
Option 3 - cap Pnl, Pnd
Anet/Ag
Pte
st/P
n
0.5 0.6 0.7 0.8 0.9 10
0.5
1
1.5
Option 4 - cap Pnl, transition Pnd
Anet/Ag
Pte
st/P
n
0.5 0.6 0.7 0.8 0.9 10
0.5
1
1.5
Option 5 - transition Pnl (Option A), transition Pnd
Anet/Ag
Pte
st/P
n
0.5 0.6 0.7 0.8 0.9 10
0.5
1
1.5
Option 6 - transition Pnl (Option B), transition Pnd
Anet/Ag
Pte
st/P
n
Figure 8.25 Test‐to‐predicted ratios for simulated local buckling‐controlled column failures as a function of net cross‐sectional area to gross cross‐sectional area
353
0.5 0.6 0.7 0.8 0.9 10
0.5
1
1.5
Option 1 - Py everywhere
Anet/Ag
Pte
st/P
n
0.5 0.6 0.7 0.8 0.9 10
0.5
1
1.5
Option 2 - Pynet everywhere
Anet/Ag
Pte
st/P
n
Distortional Controlled
0.5 0.6 0.7 0.8 0.9 10
0.5
1
1.5
Option 3 - cap Pnl, Pnd
Anet/Ag
Pte
st/P
n
0.5 0.6 0.7 0.8 0.9 10
0.5
1
1.5
Option 4 - cap Pnl, transition Pnd
Anet/Ag
Pte
st/P
n
0.5 0.6 0.7 0.8 0.9 10
0.5
1
1.5
Option 5 - transition Pnl (Option A), transition Pnd
Anet/Ag
Pte
st/P
n
0.5 0.6 0.7 0.8 0.9 10
0.5
1
1.5
Option 6 - transition Pnl (Option B), transition Pnd
Anet/Ag
Pte
st/P
n
Figure 8.26 Test‐to‐predicted ratios for simulated distortional buckling‐controlled column failures as a function of net cross‐sectional area to gross cross‐sectional
area
354
0.5 0.6 0.7 0.8 0.9 10
0.5
1
1.5
Option 1 - Py everywhere
Anet/Ag
Pte
st/P
n
0.5 0.6 0.7 0.8 0.9 10
0.5
1
1.5
Option 2 - Pynet everywhere
Anet/Ag
Pte
st/P
n
Global Controlled
0.5 0.6 0.7 0.8 0.9 10
0.5
1
1.5
Option 3 - cap Pnl, Pnd
Anet/Ag
Pte
st/P
n
0.5 0.6 0.7 0.8 0.9 10
0.5
1
1.5
Option 4 - cap Pnl, transition Pnd
Anet/Ag
Pte
st/P
n
0.5 0.6 0.7 0.8 0.9 10
0.5
1
1.5
Option 5 - transition Pnl (Option A), transition Pnd
Anet/Ag
Pte
st/P
n
0.5 0.6 0.7 0.8 0.9 10
0.5
1
1.5
Option 6 - transition Pnl (Option B), transition Pnd
Anet/Ag
Pte
st/P
n
Figure 8.27 Test‐to‐predicted ratios for simulated global buckling‐controlled column failures as a function of net cross‐sectional area to gross cross‐sectional area
355
0 5 10 15 20 25 30 35 400
0.5
1
1.5
Option 1 - Py everywhere
L/h
Pte
st/P
n
0 5 10 15 20 25 30 35 400
0.5
1
1.5
Option 2 - Pynet everywhere
L/h
Pte
st/P
n
Local Controlled
0 5 10 15 20 25 30 35 400
0.5
1
1.5
Option 3 - cap Pnl, Pnd
L/h
Pte
st/P
n
0 5 10 15 20 25 30 35 400
0.5
1
1.5
Option 4 - cap Pnl, transition Pnd
L/h
Pte
st/P
n
0 5 10 15 20 25 30 35 400
0.5
1
1.5
Option 5 - transition Pnl (Option A), transition Pnd
L/h
Pte
st/P
n
0 5 10 15 20 25 30 35 400
0.5
1
1.5
Option 6 - transition Pnl (Option B), transition Pnd
L/h
Pte
st/P
n
Figure 8.28 Test‐to‐predicted ratios for simulated local buckling‐controlled column failures as a function of column length, L, to flat web width, h
356
0 5 10 15 20 25 30 35 400
0.5
1
1.5
Option 1 - Py everywhere
L/h
Pte
st/P
n
0 5 10 15 20 25 30 35 400
0.5
1
1.5
Option 2 - Pynet everywhere
L/h
Pte
st/P
n
Distortional Controlled
0 5 10 15 20 25 30 35 400
0.5
1
1.5
Option 3 - cap Pnl, Pnd
L/h
Pte
st/P
n
0 5 10 15 20 25 30 35 400
0.5
1
1.5
Option 4 - cap Pnl, transition Pnd
L/h
Pte
st/P
n
0 5 10 15 20 25 30 35 400
0.5
1
1.5
Option 5 - transition Pnl (Option A), transition Pnd
L/h
Pte
st/P
n
0 5 10 15 20 25 30 35 400
0.5
1
1.5
Option 6 - transition Pnl (Option B), transition Pnd
L/h
Pte
st/P
n
Figure 8.29 Test‐to‐predicted ratios for simulated distortional buckling‐controlled column failures as a function of column length, L, to flat web width, h
357
0 5 10 15 20 25 30 35 400
0.5
1
1.5
Option 1 - Py everywhere
L/h
Pte
st/P
n
0 5 10 15 20 25 30 35 400
0.5
1
1.5
Option 2 - Pynet everywhere
L/h
Pte
st/P
n
Global Controlled
0 5 10 15 20 25 30 35 400
0.5
1
1.5
Option 3 - cap Pnl, Pnd
L/h
Pte
st/P
n
0 5 10 15 20 25 30 35 400
0.5
1
1.5
Option 4 - cap Pnl, transition Pnd
L/h
Pte
st/P
n
0 5 10 15 20 25 30 35 400
0.5
1
1.5
Option 5 - transition Pnl (Option A), transition Pnd
L/h
Pte
st/P
n
0 5 10 15 20 25 30 35 400
0.5
1
1.5
Option 6 - transition Pnl (Option B), transition Pnd
L/h
Pte
st/P
n
Figure 8.30 Test‐to‐predicted ratios for simulated global buckling‐controlled column failures as a function of column length, L, to web width, h
358
1.7 109BDSM comparison to experimental column database
The six DSM options are now compared to the column experiment database first
assembled in Section 1756H4.2.6.2. The database contains the elastic buckling properties of
each column, including the presence of holes and the influence of boundary conditions,
as well as the tested strengths. 1757HFigure 8.31 through 1758HFigure 8.34 compare the experiment
strengths to DSM predictions for local, distortional, and global buckling controlled
column failures. Option 1 is observed to be an accurate predictor of column strength
when local, distortional, and global slenderness are high, but overpredicts the tested
strength as slenderness decreases. This trend is consistent with the simulated
experiment comparison in Section 1759H8.1.6 and emphasizes the need for a limit on column
strength when yielding at the net section controls the failure of a column with holes.
Option 2 is even more conservative in this study when compared to the simulated
column study because the tested specimens considered only have one hole, and
therefore employing Pynet produces unrealistically high column slenderness. Option 3
shifts column specimens from the global buckling failure group to the local buckling
failure group with the Pynet limit on Pnl, resulting in improved accuracy when compared
to Option 2. Four columns in the Option 3 distortional buckling failure group are still
overpredicted by more than 10% though as observed in 1760HFigure 8.33. Option 4 and
Option 5 improve the accuracy of the underpredicted specimen strengths in Option 3
with the addition of the distortional and local yield control transitions to the net section.
Option 6 is an overly conservative predictor of columns failing by global buckling.
359
1761HTable 8.2 summarizes the test‐to‐predicted ratio statistics for all columns in the
database. Options 3, 4, and 5 are identified as the methods with the mean closest to
unity and with the lowest standard deviations. The statistics for just the stub columns
(λc<0.20) in 1762HTable 8.3 confirm the viability of DSM Options 3, 4, and 5, and provides
more direct evidence that holes limit the column strength to the net section Pynet; the
mean test‐to‐predicted ratio is 0.84 for global (yielding) failures of stub columns
employing Option 1.
Table 8.2 DSM test‐to‐predicted ratio statistics for column experiments Option Description
Mean SD φ # of tests Mean SD φ # of tests Mean SD φ # of tests1 Py everywhere 1.03 0.11 0.87 52 1.09 0.16 0.83 15 1.06 0.17 0.82 112 Pynet everywhere 1.17 0.09 0.89 47 1.22 0.13 0.87 15 1.17 0.15 0.85 163 Cap Pnl, Pnd 1.07 0.08 0.90 42 1.06 0.13 0.85 29 1.16 0.09 0.90 74 Transition Pnd, Cap Pnl 1.07 0.08 0.90 40 1.10 0.11 0.87 33 1.19 0.08 0.90 55 Transition Pnd and Pnl (Option A) 1.06 0.08 0.89 47 1.13 0.10 0.89 26 1.19 0.08 0.90 56 Transition Pnd and Pnl (Option B) 1.12 0.15 0.84 56 1.14 0.10 0.89 22 --- --- --- 0
Local buckling Distortional buckling Global buckling
Table 8.3 DSM test‐to‐predicted ratio statistics for column experiments (stub columns only)
Option DescriptionMean SD φ # of tests Mean SD φ # of tests Mean SD φ # of tests
1 Py everywhere 0.98 0.10 0.88 33 0.83 0.01 0.92 3 0.84 0.08 0.88 32 Pynet everywhere 1.12 0.07 0.90 28 1.03 0.06 0.91 3 1.07 0.12 0.86 83 Cap Pnl, Pnd 1.03 0.06 0.91 23 1.00 0.12 0.86 16 --- --- --- 04 Transition Pnd, Cap Pnl 1.04 0.06 0.91 21 1.06 0.11 0.87 18 --- --- --- 05 Transition Pnd and Pnl (Option A) 1.03 0.07 0.90 28 1.11 0.10 0.88 11 --- --- --- 06 Transition Pnd and Pnl (Option B) 1.03 0.07 0.90 29 1.11 0.10 0.88 10 --- --- --- 0
Local buckling Distortional buckling Global buckling
360
0 0.5 1 1.5 2 2.5 30
0.5
1
1.5Option 1 - Py everywhere
λl=(Pne/Pcrl
)0.5
Pn/P
ne
0 0.5 1 1.5 2 2.5 30
0.5
1
1.5Option 2 - Pynet everywhere
λl=(Pne/Pcrl
)0.5
Pn/P
ne
0 0.5 1 1.5 2 2.5 30
0.5
1
1.5Option 3 - cap Pnl
, Pnd
λl=(Pne/Pcrl
)0.5
Pn/P
ne
0 0.5 1 1.5 2 2.5 30
0.5
1
1.5Option 4 - cap Pnl
, transition Pnd
λl=(Pne/Pcrl
)0.5
Pn/P
ne
DSM Local PredictionLocal Controlled Test
0 0.5 1 1.5 2 2.5 30
0.5
1
1.5Option 5 - transition Pnl
(Option A), transition Pnd
λl=(Pne/Pcrl
)0.5
Pn/P
ne
0 0.5 1 1.5 2 2.5 30
0.5
1
1.5Option 6 - transition Pnl
(Option B), transition Pnd
λl=(Pne/Pcrl
)0.5
Pn/P
ne
Figure 8.31 Comparison of experimental test strengths to predictions for columns with local buckling‐controlled failures as a function of local slenderness (tested
strength is normalized by Pne)
361
0 0.5 1 1.5 2 2.5 30
0.5
1
1.5Option 1 - Py everywhere
λl=(Pne/Pcrl
)0.5
Pn/P
y
0 0.5 1 1.5 2 2.5 30
0.5
1
1.5Option 2 - Pynet everywhere
λl=(Pne/Pcrl
)0.5
Pn/P
y
0 0.5 1 1.5 2 2.5 30
0.5
1
1.5Option 3 - cap Pnl
, Pnd
λl=(Pne/Pcrl
)0.5
Pn/P
y
0 0.5 1 1.5 2 2.5 30
0.5
1
1.5Option 4 - cap Pnl
, transition Pnd
λl=(Pne/Pcrl
)0.5
Pn/P
y
DSM Local PredictionLocal Controlled Test
0 0.5 1 1.5 2 2.5 30
0.5
1
1.5Option 5 - transition Pnl
(Option A), transition Pnd
λl=(Pne/Pcrl
)0.5
Pn/P
y
0 0.5 1 1.5 2 2.5 30
0.5
1
1.5Option 6 - transition Pnl
(Option B), transition Pnd
λl=(Pne/Pcrl
)0.5
Pn/P
y
Figure 8.32 Comparison of experimental test strengths to predictions for columns with local buckling‐controlled failures as a function of local slenderness (tested
strength is normalized by Py)
362
0 0.5 1 1.5 2 2.5 30
0.5
1
1.5Option 1 - Py everywhere
λd=(Py/Pcrd)0.5
Pn/P
y
0 0.5 1 1.5 2 2.5 30
0.5
1
1.5Option 2 - Pynet everywhere
λdnet=(Pynet/Pcrd)0.5
Pn/P
y
DSM Dist. PredictionDist. Controlled Test
0 0.5 1 1.5 2 2.5 30
0.5
1
1.5Option 3 - cap Pnl
, Pnd
λd=(Py/Pcrd)0.5
Pn/P
y
0 0.5 1 1.5 2 2.5 30
0.5
1
1.5Option 4 - cap Pnl
, transition Pnd
λd=(Py/Pcrd)0.5
Pn/P
y
0 0.5 1 1.5 2 2.5 30
0.5
1
1.5Option 5 - transition Pnl
(Option A), transition Pnd
λd=(Py/Pcrd)0.5
Pn/P
y
0 0.5 1 1.5 2 2.5 30
0.5
1
1.5Option 6 - transition Pnl
(Option B), transition Pnd
λd=(Py/Pcrd)0.5
Pn/P
y
Figure 8.33 Comparison of experimental test strengths to predictions for columns with distortional buckling‐controlled failures as a function of distortional
slenderness
363
0 0.5 1 1.5 2 2.5 30
0.5
1
1.5
λc=(Py/Pcre)0.5
Pn/P
y
Option 1 - Py everywhere
0 0.5 1 1.5 2 2.5 30
0.5
1
1.5
λcnet=(Pynet/Pcre)0.5
Pn/P
y
Option 2 - Pynet everywhere DSM Global Prediction
Global Controlled Test
0 0.5 1 1.5 2 2.5 30
0.5
1
1.5
λc=(Py/Pcre)0.5
Pn/P
y
Option 3 - cap Pnl, Pnd
0 0.5 1 1.5 2 2.5 30
0.5
1
1.5
λc=(Py/Pcre)0.5
Pn/P
y
Option 4 - cap Pnl, transition Pnd
0 0.5 1 1.5 2 2.5 30
0.5
1
1.5
λc=(Py/Pcre)0.5
Pn/P
y
Option 5 - transition Pnl (Option A), transition Pnd
0 0.5 1 1.5 2 2.5 30
0.5
1
1.5
λc=(Py/Pcre)0.5
Pn/P
y
Option 6 - transition Pnl (Option B), transition Pnd
Figure 8.34 Comparison of experimental test strengths to predictions for columns with global buckling‐controlled failures as a function of global slenderness
364
0 0.5 1 1.5 2 2.5 30
0.5
1
1.5
Option 1 - Py everywhere
λl=(Pne/Pcrl
)0.5
Pte
st/P
n
0 0.5 1 1.5 2 2.5 30
0.5
1
1.5
Option 2 - Pynet everywhere
λl=(Pne/Pcrl
)0.5
Pte
st/P
n
Local Controlled
0 0.5 1 1.5 2 2.5 30
0.5
1
1.5
Option 3 - cap Pnl, Pnd
λl=(Pne/Pcrl
)0.5
Pte
st/P
n
0 0.5 1 1.5 2 2.5 30
0.5
1
1.5
Option 4 - cap Pnl, transition Pnd
λl=(Pne/Pcrl
)0.5
Pte
st/P
n
0 0.5 1 1.5 2 2.5 30
0.5
1
1.5
Option 5 - transition Pnl (Option A), transition Pnd
λl=(Pne/Pcrl
)0.5
Pte
st/P
n
0 0.5 1 1.5 2 2.5 30
0.5
1
1.5
Option 6 - transition Pnl (Option B), transition Pnd
λl=(Pne/Pcrl
)0.5
Pte
st/P
n
Figure 8.35 Test‐to‐predicted ratios for experiment local buckling‐controlled column failures as a function of local slenderness
365
0 0.5 1 1.5 2 2.5 30
0.5
1
1.5
Option 1 - Py everywhere
λd=(Py/Pcrd)0.5
Pte
st/P
n
0 0.5 1 1.5 2 2.5 30
0.5
1
1.5
Option 2 - Pynet everywhere
λdnet=(Pynet/Pcrd)0.5
Pte
st/P
n
Distortional Controlled
0 0.5 1 1.5 2 2.5 30
0.5
1
1.5
Option 3 - cap Pnl, Pnd
λd=(Py/Pcrd)0.5
Pte
st/P
n
0 0.5 1 1.5 2 2.5 30
0.5
1
1.5
Option 4 - cap Pnl, transition Pnd
λd=(Py/Pcrd)0.5
Pte
st/P
n
0 0.5 1 1.5 2 2.5 30
0.5
1
1.5
Option 5 - transition Pnl (Option A), transition Pnd
λd=(Py/Pcrd)0.5
Pte
st/P
n
0 0.5 1 1.5 2 2.5 30
0.5
1
1.5
Option 6 - transition Pnl (Option B), transition Pnd
λd=(Py/Pcrd)0.5
Pte
st/P
n
Figure 8.36 Test‐to‐predicted ratios for experiment distortional buckling‐controlled column failures as a function of distortional slenderness
366
0 0.5 1 1.5 2 2.5 30
0.5
1
1.5
Option 1 - Py everywhere
λc=(Py/Pcre)0.5
Pte
st/P
ne
0 0.5 1 1.5 2 2.5 30
0.5
1
1.5
Option 2 - Pynet everywhere
λcnet=(Pynet/Pcre)0.5
Pte
st/P
ne
Global Controlled
0 0.5 1 1.5 2 2.5 30
0.5
1
1.5
Option 3 - cap Pnl, Pnd
λc=(Py/Pcre)0.5
Pte
st/P
ne
0 0.5 1 1.5 2 2.5 30
0.5
1
1.5
Option 4 - cap Pnl, transition Pnd
λc=(Py/Pcre)0.5
Pte
st/P
ne
0 0.5 1 1.5 2 2.5 30
0.5
1
1.5
Option 5 - transition Pnl (Option A), transition Pnd
λc=(Py/Pcre)0.5
Pte
st/P
ne
0 0.5 1 1.5 2 2.5 30
0.5
1
1.5
Option 6 - transition Pnl (Option B), transition Pnd
λc=(Py/Pcre)0.5
Pte
st/P
ne
Figure 8.37 Test‐to‐predicted ratios for experiment global buckling‐controlled column failures as a function of global slenderness
367
0.5 0.6 0.7 0.8 0.9 10
0.5
1
1.5
Option 1 - Py everywhere
Anet/Ag
Pte
st/P
n
0.5 0.6 0.7 0.8 0.9 10
0.5
1
1.5
Option 2 - Pynet everywhere
Anet/Ag
Pte
st/P
n
Local Controlled
0.5 0.6 0.7 0.8 0.9 10
0.5
1
1.5
Option 3 - cap Pnl, Pnd
Anet/Ag
Pte
st/P
n
0.5 0.6 0.7 0.8 0.9 10
0.5
1
1.5
Option 4 - cap Pnl, transition Pnd
Anet/Ag
Pte
st/P
n
0.5 0.6 0.7 0.8 0.9 10
0.5
1
1.5
Option 5 - transition Pnl (Option A), transition Pnd
Anet/Ag
Pte
st/P
n
0.5 0.6 0.7 0.8 0.9 10
0.5
1
1.5
Option 6 - transition Pnl (Option B), transition Pnd
Anet/Ag
Pte
st/P
n
Figure 8.38 Test‐to‐predicted ratios for experiment local buckling‐controlled column failures as a function of net cross‐sectional area to gross cross‐sectional area
368
0.5 0.6 0.7 0.8 0.9 10
0.5
1
1.5
Option 1 - Py everywhere
Anet/Ag
Pte
st/P
n
0.5 0.6 0.7 0.8 0.9 10
0.5
1
1.5
Option 2 - Pynet everywhere
Anet/Ag
Pte
st/P
n
Distortional Controlled
0.5 0.6 0.7 0.8 0.9 10
0.5
1
1.5
Option 3 - cap Pnl, Pnd
Anet/Ag
Pte
st/P
n
0.5 0.6 0.7 0.8 0.9 10
0.5
1
1.5
Option 4 - cap Pnl, transition Pnd
Anet/Ag
Pte
st/P
n
0.5 0.6 0.7 0.8 0.9 10
0.5
1
1.5
Option 5 - transition Pnl (Option A), transition Pnd
Anet/Ag
Pte
st/P
n
0.5 0.6 0.7 0.8 0.9 10
0.5
1
1.5
Option 6 - transition Pnl (Option B), transition Pnd
Anet/Ag
Pte
st/P
n
Figure 8.39 Test‐to‐predicted ratios for experiment distortional buckling‐controlled column failures as a function of net cross‐sectional area to gross cross‐sectional
area
369
0.5 0.6 0.7 0.8 0.9 10
0.5
1
1.5
Option 1 - Py everywhere
Anet/Ag
Pte
st/P
n
0.5 0.6 0.7 0.8 0.9 10
0.5
1
1.5
Option 2 - Pynet everywhere
Anet/Ag
Pte
st/P
n
Global Controlled
0.5 0.6 0.7 0.8 0.9 10
0.5
1
1.5
Option 3 - cap Pnl, Pnd
Anet/Ag
Pte
st/P
n
0.5 0.6 0.7 0.8 0.9 10
0.5
1
1.5
Option 4 - cap Pnl, transition Pnd
Anet/Ag
Pte
st/P
n
0.5 0.6 0.7 0.8 0.9 10
0.5
1
1.5
Option 5 - transition Pnl (Option A), transition Pnd
Anet/Ag
Pte
st/P
n
0.5 0.6 0.7 0.8 0.9 10
0.5
1
1.5
Option 6 - transition Pnl (Option B), transition Pnd
Anet/Ag
Pte
st/P
n
Figure 8.40 Test‐to‐predicted ratios for experiment global buckling‐controlled column failures as a function of net cross‐sectional area to gross cross‐sectional area
370
0 5 10 15 20 25 30 35 400
0.5
1
1.5
Option 1 - Py everywhere
L/h
Pte
st/P
n
0 5 10 15 20 25 30 35 400
0.5
1
1.5
Option 2 - Pynet everywhere
L/h
Pte
st/P
n
Local Controlled
0 5 10 15 20 25 30 35 400
0.5
1
1.5
Option 3 - cap Pnl, Pnd
L/h
Pte
st/P
n
0 5 10 15 20 25 30 35 400
0.5
1
1.5
Option 4 - cap Pnl, transition Pnd
L/h
Pte
st/P
n
0 5 10 15 20 25 30 35 400
0.5
1
1.5
Option 5 - transition Pnl (Option A), transition Pnd
L/h
Pte
st/P
n
0 5 10 15 20 25 30 35 400
0.5
1
1.5
Option 6 - transition Pnl (Option B), transition Pnd
L/h
Pte
st/P
n
Figure 8.41 Test‐to‐predicted ratios for experiment local buckling‐controlled column failures as a function of column length, L, to flat web width, h
371
0 5 10 15 20 25 30 35 400
0.5
1
1.5
Option 1 - Py everywhere
L/h
Pte
st/P
n
0 5 10 15 20 25 30 35 400
0.5
1
1.5
Option 2 - Pynet everywhere
L/h
Pte
st/P
n
Distortional Controlled
0 5 10 15 20 25 30 35 400
0.5
1
1.5
Option 3 - cap Pnl, Pnd
L/h
Pte
st/P
n
0 5 10 15 20 25 30 35 400
0.5
1
1.5
Option 4 - cap Pnl, transition Pnd
L/h
Pte
st/P
n
0 5 10 15 20 25 30 35 400
0.5
1
1.5
Option 5 - transition Pnl (Option A), transition Pnd
L/h
Pte
st/P
n
0 5 10 15 20 25 30 35 400
0.5
1
1.5
Option 6 - transition Pnl (Option B), transition Pnd
L/h
Pte
st/P
n
Figure 8.42 Test‐to‐predicted ratios for experiment distortional buckling‐controlled column failures as a function of column length, L, to flat web width, h
372
0 5 10 15 20 25 30 35 400
0.5
1
1.5
Option 1 - Py everywhere
L/h
Pte
st/P
n
0 5 10 15 20 25 30 35 400
0.5
1
1.5
Option 2 - Pynet everywhere
L/h
Pte
st/P
n
Global Controlled
0 5 10 15 20 25 30 35 400
0.5
1
1.5
Option 3 - cap Pnl, Pnd
L/h
Pte
st/P
n
0 5 10 15 20 25 30 35 400
0.5
1
1.5
Option 4 - cap Pnl, transition Pnd
L/h
Pte
st/P
n
0 5 10 15 20 25 30 35 400
0.5
1
1.5
Option 5 - transition Pnl (Option A), transition Pnd
L/h
Pte
st/P
n
0 5 10 15 20 25 30 35 400
0.5
1
1.5
Option 6 - transition Pnl (Option B), transition Pnd
L/h
Pte
st/P
n
Figure 8.43 Test‐to‐predicted ratios for experiment global buckling‐controlled column failures as a function of column length, L, to web width, h
373
1.8 110BRecommendations – DSM for columns with holes
Options 3, 4, and 5 are presented as viable proposals for extending DSM to columns
with holes. This recommendation is based on the test‐to‐predicted statistics and data
trends presented in Section 1763H8.1.6 and Section 1764H8.1.7, and also considers the effort to
implement the modifications and their ease of use by design engineers. Option 3
accounts for the reduction in column strength from the presence of holes by capping Pnl
and Pnd at Pynet. This is a simple modification to implement in the Specification and
avoids additional calculation work for a design engineer (except for that required to
calculate the critical elastic buckling loads including the influence of the holes). Options
4 and 5 are refinements of Option 3, where the cap on Pnl and Pnd becomes a transition
from an elastic buckling controlled failure mode to a yield controlled failure at Pynet.
These two methods require additional effort from the designer when compared to
Option 3, but they have an important advantage. Options 4 and 5 are more closely tied
to the failure mechanisms influencing column strength because they capture the yield
transition to the net section in their predictions. The transitions increase the probability
that strength will be accurately predicted for general column and hole geometries.
Option 5 has the additional advantage of capturing the influence of a yield transition for
closed cross‐sections that do not experience distortional buckling. This generality is
what motivates the use of the Direct Strength Method (AISI‐S100 2007, Appendix 1).
374
8.2 49BDSM for laterally braced beams with holes
2.1 111BDatabase of simulated column experiments
Simulated experiments were conducted on 125 C‐section laterally braced beams with
evenly‐spaced circular web holes in ABAQUS. Cross‐sections were specifically selected
with custom Matlab code employing the existing DSM design curves to identify beams
predisposed to local and distortional buckling‐controlled failures. The cross‐sections
were chosen from a catalog of 99 industry standard C‐sections published by the Steel
Stud Manufacturers Association (SSMA 2001). The nominal out‐to‐out dimensions
provided in the SSMA catalog were converted to centerline dimensions and then
constructed in ABAQUS with the meshing procedure described in Section 1765H7.2.1.1. The
beams in the database have a constant length L=48 in. to accommodate multiple local
and distortional buckling half‐waves along the beam. Evenly spaced circular web holes
were placed in the columns with hole spacing S (defined in 1766HFigure 3.2) of 16 inches (i.e.,
three evenly spaced holes). The holes were centered transversely in the web and their
depth (diameter), hhole, was varied such that the ratio of the net moment of inertia, Inet, to
the gross cross‐sectional area, Ig, ranged between 0.85 and 1.0.
The ABAQUS boundary conditions and application of loading, described in 1767HFigure
8.44, are implemented to be consistent with CUFSM, i.e. pinned‐pinned and free‐to‐
warp with a uniform stress applied at the member ends. Each beam is laterally braced
by restraining the compression flange at the midlength of the beam. (Initial modeling
trials, where all nodes centered in the compression flange were laterally restrained,
375
resulted in simulated strengths 25% higher than DSM predictions for beams without
holes.) Consistent nodal loads were applied to simulate the linear stress gradient at the
beam ends (see Section 1768H7.2.1.2 for information on S9R5 consistent nodal loads). The
loads (a reference moment of 1 kip∙in. was applied at each end in ABAQUS) were
distributed over the first two sets of cross‐section nodes to avoid localized failures at the
loaded edges.
2
1
3
End cross-section nodes restrained in 2 and 3
54
6
End cross-section nodes restrained in 2 and 3
Node centered in compression flange at longitudinal midline restrained in 1 (to prevent rigid body motion) and 3 (for laterally bracing)
Moment applied as consistent nodal loads over two sets of cross-section nodes to avoid edge failures (Typ.)
Figure 8.44 ABAQUS simulated beam experiments boundary conditions and application of loading
The ABAQUS simulations were performed with the modified Riks nonlinear
solution algorithm. Automatic time stepping was enabled with a suggested initial arc
length step of 1 (the Riks method increments in units of energy, in this case kip∙in.), a
maximum step size of 3, and the maximum number of solution increments set at 300.
Metal plasticity was simulated with the material modeling procedure described in
Section 1769H7.2.1.4. The plastic true stress‐strain curve for specimen 362‐1‐48‐H in 1770HAppendix
H was assumed for all column models (but modified such that plasticity starts at the
yield stress, see Section 1771H7.2.1.4), where the steel yield stress Fy=58.6 ksi. Residual stresses
376
and initial plastic strains, as discussed in Section 1772H7.2.1.6, were not considered in the
ABAQUS models because their implementation requires further validation and they
were not observed to markedly influence column ultimate strength (see 1773HFigure 7.48 and
1774HFigure 7.49).
Imperfections were imposed on the initial beam geometry in ABAQUS with custom
Matlab code which combines the local and distortional buckling cross‐section mode
shapes from CUFSM along the column length. Two simulations were performed for
each beam, one model with 25% CDF local and distortional imperfection magnitudes
and the other model with 75% CDF local and distortional imperfection magnitudes (see
Section 1775H7.2.1.5 for local and distortional imperfection definitions).
The local (Mcrl) and distortional (Mcrd) critical elastic buckling loads were predicted
for each beam with custom Matlab code based on the CUFSM prediction methods
described in Section 1776H4.3. The database of simulated beam experiments, including cross‐
section type, column and hole geometry, simulated ultimate strength (Mtest25 and Mtest75)
and critical elastic buckling loads for each beam (including the presence of holes) is
provided in 1777HAppendix L.
2.2 112BLocal buckling study
Twelve beams from the simulation database in 1778HAppendix L were chosen to study
the influence of web holes on the ultimate strength of laterally braced beams
predisposed to a local buckling‐controlled failure. The beams have SSMA cross‐sections
which result in a local buckling slenderness, λl, ranging from 1.3 to 2.0. (The slenderness
377
range considered here is relatively narrow because only 12 of the 99 SSMA cross‐
sections, when employed as laterally braced beams, are controlled by a local buckling
failure. The majority of beam cross‐sections are predicted to exhibit a distortional
buckling‐controlled failure.) The web of each beam contains three evenly spaced
circular holes where the hole spacing S=16 in. The hole depth (diameter), hhole, is varied
for each beam to produce Inet/Ig of 1.0 (no holes), 0.95, 0.90, and 0.85. Refer to 1779HAppendix
L, Study Type L, for specific beam cross‐section and hole geometry information.
The simulation results for Inet/Ig =1.0, 0.95, 0.90, and 0.85, are compared to the DSM
distortional buckling prediction curve in 1780HFigure 8.46 to 1781HFigure 8.49. The beam strengths,
Mtest25 and Mtest75, without holes (Inet/Ig =1.0) are consistent with the DSM design curve as
shown in 1782HFigure 8.46a, confirming that the nonlinear simulation protocol developed for
columns in Section 1783H7.2 is also viable when conducting cold‐formed steel beam
simulations. The mean and standard deviation of the simulated test to predicted ratio is
1.05 and 0.05 respectively for 25% CDF local and distortional imperfections, and 1.03
and 0.05 for 75% CDF local and distortional imperfections. For the beams with holes,
the simulated test strengths diverge from the DSM prediction curve as local slenderness,
λl=(Myg/Mcrl)0.5, decreases as shown in 1784HFigure 8.47a to 1785HFigure 8.49a (Myg is the yield
moment of the column calculated with the gross cross‐sectional area Ig). This divergent
trend in Mtest with decreasing λl is consistent with the column results with holes
discussed in Section 1786H8.1, where elastic buckling controlled the failure when slenderness
was high and transitioned to yielding and collapse of the net section as slenderness
decreased. 1787HFigure 8.45 shows the load‐deformation response at ultimate limit state for
378
an SSMA 800S162‐43 beam considered in this study, and highlights the transition from
an elastic buckling controlled‐failure to a yield controlled‐failure at the net section as
hole size increases.
1.0 0.95 0.90 0.85Inet/Ig0.72 0.68 0.64 0.61 Mtest25/Myg
1.45 1.63 1.45 1.45λl
Elastic buckling controlled failure Yielding and collapse of net section
Figure 8.45 SSMA 800S162‐43 beam with web holes considered in the DSM local buckling study
Two modification options are proposed for the DSM local buckling beam design
curve:
Local Buckling (Option A) The nominal flexural strength, Mnl, for local buckling shall be calculated in accordance with the following:
(a) For λl ≤λl1 Mnl = Mne≤ Mynet (cap on beam strength)
(b) For λl1<λl≤λl2
Mnl = ( ) ⎟⎟⎠
⎞⎜⎜⎝
⎛−−
⎟⎟⎠
⎞⎜⎜⎝
⎛−−
12
122ynetynet MMM
ll
ll
l
ll λλ
λλλλ (nonlinear yield transition when Mynet/Mne ≤1)
(c) For λl > λl2
Mnl = ne
4.0
ne
cr
4.0
ne
cr MMM
MM
15.01 ⎟⎟⎠
⎞⎜⎜⎝
⎛⎟⎟
⎠
⎞
⎜⎜
⎝
⎛⎟⎟⎠
⎞⎜⎜⎝
⎛− ll (DSM local buckling curve, unchanged)
where λl =
lcrne MM
λl1 = )MM(776.0 neynet≤0.776
λl2 = ( )( )4.1MM4.2776.0 5.3neynet −− , Mynet/Mne≤1
= 0.776, Mynet/Mne>1 (no transition when Mynet/Mne >1 ) Μl2 = ( )( )( ) ne
8.02
8.02 M1115.01 ll λλ−
Mynet = SfnetFy≥0.80My (limit reduction of the net section to 0.8My)
Sfnet = Section modulus at the hole(s) referenced to the extreme fiber at first yield Mcrl = Critical elastic local beam buckling load including hole(s)
379
Local Buckling (Option B) The nominal axial strength, Mnl, for local buckling shall be calculated in accordance with the following:
(a) For λl ≤λl1 Mnl = Mynet (Mne /My) (cap on column strength)
(b) For λl1<λl≤λl2
Mnl = ⎟⎟⎠
⎞⎜⎜⎝
⎛−−
⎟⎟⎠
⎞⎜⎜⎝
⎛⎟⎟⎠
⎞⎜⎜⎝
⎛−⎟
⎟⎠
⎞⎜⎜⎝
⎛−⎟
⎟⎠
⎞⎜⎜⎝
⎛
12
122
y
neynet
y
neynet M
MMM
MMM
ll
ll
l
ll λλ
λλλλ (nonlinear yield transition)
(c) For λl > λl2
Mnl = ne
4.0
ne
cr
4.0
ne
cr MMM
MM
15.01 ⎟⎟⎠
⎞⎜⎜⎝
⎛⎟⎟
⎠
⎞
⎜⎜
⎝
⎛⎟⎟⎠
⎞⎜⎜⎝
⎛− ll (DSM local buckling curve, unchanged)
where λl =
lcrne MM
λl1 = )MM(776.0 yynet≤0.776
λl2 = ( )( )4.1MM4.2776.0 5.3yynet −−
Μl2 = ( )( )( ) ne8.0
28.0
2 M1115.01 ll λλ− Mynet = SfnetFy≥0.80My (limit reduction of the net section to 0.8My)
Sfnet = Section modulus at the hole(s) referenced to the extreme fiber at first yield Mcrl = Critical elastic local beam buckling load including hole(s)
The framework for Option A and Option B is based on the proposed modifications
to the DSM local buckling column design curve presented in Section 1788H8.1.4. Option A
imposes a transition from the DSM local buckling curve to the net section limit, Mynet,
when Mynet< Mne. When Mynet>Mne, Option A assumes that holes influence only the critical
elastic buckling loads (Mcrl, Mcre) but otherwise do not change the failure mode of the
beam. Option B also imposes a transition to the net beam strength from the DSM local
failure curve, although in this case the yield transition occurs for all values of Mynet/Mne.
The proposed transition from the elastic buckling failure regime to the yield plateau is
nonlinear for both Options A and B as demonstrated in 1789HFigure 8.47a to 1790HFigure 8.49a, in
contrast to the linear transition for cold‐formed steel columns with holes (see Section
1791H8.1.4).
380
All beams considered in this study are laterally braced, i.e. global (lateral‐torsional)
buckling does not influence beam strength, and therefore Option A and B will produce
the same strength predictions. The validity of both options for laterally braced beams is
evaluated in the following section with the simulation database in 1792HAppendix L and the
experiment database assembled in 1793HChapter 4. Future work is planned to evaluate
Option A and B for unbraced cold‐formed steel beams with holes, where lateral‐
torsional buckling influences beam strength.
0 0.5 1 1.5 2 2.5 3
0
0.2
0.4
0.6
0.8
1
1.2
1.4
Mte
st/M
n l
local slenderness, λl=(My/Mcrl
)0.5
DSM (no holes)FE 25% CDF imperfectionsFE 75% CDF imperfections
0 0.5 1 1.5 2 2.5 3
0
0.2
0.4
0.6
0.8
1
1.2
1.4
Mte
st/M
n l
local slenderness, λl=(My/Mcrl
)0.5
DSM (no holes)DSM (proposed, with holes)FE 25% CDF imperfectionsFE 75% CDF imperfections
Figure 8.46 Comparison of simulated beam strengths (Inet/Ig=1.0, no holes) to (a) the existing DSM local buckling design curve and to (b) the proposed DSM local buckling curve for beams with holes
0 0.5 1 1.5 2 2.5 30
0.2
0.4
0.6
0.8
1
1.2
1.4
Mte
st/M
n l
local slenderness, λl=(My/Mcrl
)0.5
DSM (no holes)FE 25% CDF imperfectionsFE 75% CDF imperfections
0 0.5 1 1.5 2 2.5 3
0
0.2
0.4
0.6
0.8
1
1.2
1.4
Mte
st/M
n l
local slenderness, λl=(My/Mcrl
)0.5
DSM (no holes)DSM (proposed, with holes)FE 25% CDF imperfectionsFE 75% CDF imperfections
Figure 8.47 Comparison of simulated beam strengths (Inet/Ig=0.95) to (a) the existing DSM local buckling design curve and to (b) the proposed DSM local buckling curve for beams with holes
381
0 0.5 1 1.5 2 2.5 30
0.2
0.4
0.6
0.8
1
1.2
1.4
Mte
st/M
n l
local slenderness, λl=(My/Mcrl
)0.5
DSM (no holes)FE 25% CDF imperfectionsFE 75% CDF imperfections
0 0.5 1 1.5 2 2.5 30
0.2
0.4
0.6
0.8
1
1.2
1.4
Mte
st/M
n l
local slenderness, λl=(My/Mcrl
)0.5
DSM (no holes)DSM (proposed, with holes)FE 25% CDF imperfectionsFE 75% CDF imperfections
Figure 8.48 Comparison of simulated beam strengths (Inet/Ig=0.90) to (a) the existing DSM local buckling design curve and to (b) the proposed DSM local buckling curve for beams with holes
0 0.5 1 1.5 2 2.5 30
0.2
0.4
0.6
0.8
1
1.2
1.4
Mte
st/M
n l
local slenderness, λl=(My/Mcrl
)0.5
DSM (no holes)FE 25% CDF imperfectionsFE 75% CDF imperfections
0 0.5 1 1.5 2 2.5 30
0.2
0.4
0.6
0.8
1
1.2
1.4
Mte
st/M
n l
local slenderness, λl=(My/Mcrl
)0.5
DSM (no holes)DSM (proposed, with holes)FE 25% CDF imperfectionsFE 75% CDF imperfections
Figure 8.49 Comparison of simulated beam strengths (Inet/Ig=0.85) to (a) the existing DSM local buckling design curve and to (b) the proposed DSM local buckling curve for beams with holes
2.3 113BDistortional buckling study
A group of 11 beams from the SSMA beam simulation database was chosen to
evaluate the influence of the ratio Inet/Ig on the tested strength of beams predicted to
collapse with a distortional failure mode. (Ig is the gross moment of inertia of a beam
and Inet is the moment of inertia at the location of a hole.) The beams have SSMA cross‐
sections which result in a distortional buckling slenderness, λd, ranging from 0.6 to 1.6.
(All SSMA cross‐sections, employed as beams and controlled by a distortional buckling
failure, lie within this slenderness range.) In this study the beam depths range from 4 in.
382
to 12 in. The web of each beam has three circular holes where the hole spacing S=16 in
(see 1794HFigure 3.2 for the definition of S). The hole depth (diameter), hhole, is varied for each
beam to produce Inet/Ig of 1.0 (no holes), 0.95, and 0.90. Refer to 1795HAppendix L, Study Type
D, for specific cross‐section and hole geometry information.
The simulation results for Inet/Ig =1.0, 0.95, and 0.90 are compared to the DSM
distortional buckling prediction curve in 1796HFigure 8.51 to 1797HFigure 8.53. The beam strengths,
Mtest25 and Mtest75, without holes (Inet/Ig =1.0) are consistent with the DSM distortional
buckling design curve as shown in 1798HFigure 8.51a, with a trend of increasingly
conservative predictions as distortional slenderness increases. The mean and standard
deviation of the simulated test to predicted ratio is 1.08 and 0.08 respectively for 25%
CDF local and distortional imperfections, and 1.02 and 0.12 for 75% CDF imperfections.
For the beams with holes, the simulated test strengths demonstrate a slight divergence
from the DSM prediction curve as distortional slenderness, λd=(Myg/Mcrd)0.5, decreases as
shown in 1799HFigure 8.52a and 1800HFigure 8.53a (Myg is the yield moment of the beam calculated
with the gross cross‐sectional area Ig). ( 1801HFigure 8.52a and 1802HFigure 8.53a also demonstrate
that Mcrd, predicted with the simplified method in Section 1803H4.3.2.2, increases distortional
slenderness and shifts the simulated data off of the prediction curve. Future research is
planned to improve the accuracy of this simplified method.) This divergent trend in
Mtest was also observed in the local buckling‐controlled beam study in Section 1804H8.2.2 and
the column studies presented in Section 1805H8.1. As λd decreases, the beam failure mode
transitions from a distortional buckling failure to yielding and collapse of the net
section. 1806HFigure 8.50 highlights this transition for the SSMA 550S162‐54 beam considered
383
in this study by comparing the deformed shape at ultimate limit state as hole size
increases.
1.0 0.95 0.90Inet/Ig0.90 0.89 0.83Mtest25/Myg
0.87 0.92 0.93λd
Elastic buckling controlled failure Yielding and collapse of net section
Figure 8.50 SSMA 550S162‐54 structural stud failure mode transition from distortional buckling to yielding at the net section
The observations from this study are used to formulate a modified DSM distortional
curve for beams with holes which captures the failure mechanism transition from
yielding at the net cross‐section to a distortional type failure mode and limits the
strength of the beam to the yield moment at the net section:
Distortional Buckling The nominal flexural strength, Mnd, for distortional buckling shall be calculated in accordance with the following:
(a) For λd ≤λd1 Mnd = ynetM (cap on column strength) (b) For λd1<λd≤λd2
Mnd = d
1d2d
2dynetynet
MMM λ
λλ ⎟⎟⎠
⎞⎜⎜⎝
⎛−
−− (yield control transition)
(c) For λd > λd2
Mnd = y
5.0
y
crd
5.0
y
crd MMM
MM
22.01 ⎟⎟⎠
⎞⎜⎜⎝
⎛⎟⎟
⎠
⎞
⎜⎜
⎝
⎛
⎟⎟⎠
⎞⎜⎜⎝
⎛− (existing DSM distortional curve)
where λd =
crdy MM
λd1 = )MM(673.0 yynet
λd2 = ( )( )7.0MM7.1673.0 7.1yynet −−
Μd2 = ( )( )( ) y2d2d M1122.01 λλ−
Mynet = SfnetFy≥0.80My (limit reduction of the net section to 0.8My)
Sfnet = Section modulus at the hole(s) referenced to the extreme fiber at first yield Mcrd = Critical elastic distortional beam buckling load including hole(s)
384
The modified DSM distortional curve is added in 1807HFigure 8.51b to 1808HFigure 8.53b as Inet/Ig
decreases, simulating the transition from the existing DSM curve to the net section
strength limit exhibited by the simulated test data. The linear portion of the modified
prediction curve represents the unstiffened strip distortional collapse mechanism and
the nonlinear portion represents a collapse mechanism driven by distortional buckling.
This proposed modification to the DSM distortional prediction curve will be compared
against the beam experiments database developed in Section 1809H4.3.1 as a part of several
proposed DSM options considered later in this chapter.
0 0.5 1 1.5 2 2.5 30
0.2
0.4
0.6
0.8
1
1.2
1.4
distortional slenderness, λd=(My/Mcrd)0.5
Mte
st/M
y
DSM (no hole)FE 25% CDF imperfectionsFE 75% CDF imperfections
0 0.5 1 1.5 2 2.5 3
0
0.2
0.4
0.6
0.8
1
1.2
1.4
distortional slenderness, λd=(My/Mcrd)0.5
Mte
st/M
y
DSM (no hole)DSM (proposed, with holes)FE 25% CDF imperfectionsFE 75% CDF imperfections
Figure 8.51 Comparison of simulated beam strengths (Inet/Ig=1.0) to (a) the existing DSM distortional buckling design curve and to (b) the proposed DSM distortional buckling curve for beams with holes
0 0.5 1 1.5 2 2.5 30
0.2
0.4
0.6
0.8
1
1.2
1.4
distortional slenderness, λd=(My/Mcrd)0.5
Mte
st/M
y
DSM (no hole)FE 25% CDF imperfectionsFE 75% CDF imperfections
0 0.5 1 1.5 2 2.5 3
0
0.2
0.4
0.6
0.8
1
1.2
1.4
distortional slenderness, λd=(My/Mcrd)0.5
Mte
st/M
y
DSM (no hole)DSM (proposed, with holes)FE 25% CDF imperfectionsFE 75% CDF imperfections
Figure 8.52 Comparison of simulated beam strengths (Inet/Ig=0.95) to (a) the existing DSM distortional buckling design curve and to (b) the proposed DSM distortional buckling curve for beams with holes
385
0 0.5 1 1.5 2 2.5 30
0.2
0.4
0.6
0.8
1
1.2
1.4
distortional slenderness, λd=(My/Mcrd)0.5
Mte
st/M
y
DSM (no hole)FE 25% CDF imperfectionsFE 75% CDF imperfections
0 0.5 1 1.5 2 2.5 3
0
0.2
0.4
0.6
0.8
1
1.2
1.4
distortional slenderness, λd=(My/Mcrd)0.5
Mte
st/M
y
DSM (no hole)DSM (proposed, with holes)FE 25% CDF imperfectionsFE 75% CDF imperfections
Figure 8.53 Comparison of simulated beam strengths (Inet/Ig=0.90) to (a) the existing DSM distortional buckling design curve and to (b) the proposed DSM distortional buckling curve for beams with holes
2.4 114BPresentation and evaluation of DSM options
Six options for extending DSM to laterally braced beams with holes are evaluated in
this section. The options range from simple substitutions in the existing code to more
involved modifications, including the incorporation of the design curve transitions
discussed in Section 1810H8.2.2 and Section 1811H8.2.3 for local and distortional buckling.
386
2.4.1 183BDescription of DSM options
Option 1: Include hole(s) in Mcr determinations, ignore hole otherwise This method, in presentation, appears identical to currently available DSM expressions
Lateral-Torsional Buckling The nominal flexural strength, Mne, for lateral-torsional buckling shall be calculated in accordance with the following:
(a) for Mcre < 0.56 My Mne = Mcre
(b) for 2.78My ≥ Mcre ≥ 0.56My ⎟⎟⎠
⎞⎜⎜⎝
⎛−=
cre
yyne M36
M101M
910M
(c) for Mcre >2.78My Mne= My
where Mcre= Critical elastic global beam buckling load … (including hole(s))
Local Buckling The nominal flexural strength, Mnl, for local buckling shall be calculated in accordance with the following:
(a) For λl ≤0.776 Mnl = Mne
(b) For λl > 0.776
Mnl = ne
4.0
ne
cr
4.0
ne
cr MMM
MM
15.01 ⎟⎟⎠
⎞⎜⎜⎝
⎛⎟⎟
⎠
⎞
⎜⎜
⎝
⎛⎟⎟⎠
⎞⎜⎜⎝
⎛− ll
where λl =
lcrne MM
Mcrl = Critical elastic local beam buckling load including hole(s) Mne= defined in section above
Distortional Buckling The nominal flexural strength, Mnd, for distortional buckling shall be calculated in accordance with the following:
(a) For λd ≤0.673 Mnd = yM (b) For λd >0.673
Mnd = y
5.0
y
crd
5.0
y
crd MMM
MM
22.01 ⎟⎟⎠
⎞⎜⎜⎝
⎛⎟⎟
⎠
⎞
⎜⎜
⎝
⎛
⎟⎟⎠
⎞⎜⎜⎝
⎛−
where λd =
crdy MM
Mcrd = Critical elastic distortional beam buckling load including hole(s)
387
Option 2: Include hole(s) in Mcr determinations, Use Mynet everywhere The only change in this method is to replace My with Mynet
Lateral-Torsional Buckling The nominal flexural strength, Mne, for lateral-torsional buckling shall be calculated in accordance with the following:
(a) for Mcre < 0.56 Mynet Mne = Mcre
(b) for 2.78Mynet ≥ Mcre ≥ 0.56Mynet ⎟⎟⎠
⎞⎜⎜⎝
⎛−=
cre
ynetynetne M36
M101M
910M
(c) for Mcre >2.78Mynet Mne= Mynet
where Mcre= Critical elastic global beam buckling load … (including hole(s))
Mynet = SfnetFy≥0.80My
Sfnet = Section modulus at the hole(s) referenced to the extreme fiber at first yield
Local Buckling The nominal flexural strength, Mnl, for local buckling shall be calculated in accordance with the following:
(a) For λl ≤0.776 Mnl = Mne
(b) For λl > 0.776
Mnl = ne
4.0
ne
cr
4.0
ne
cr MMM
MM
15.01 ⎟⎟⎠
⎞⎜⎜⎝
⎛⎟⎟
⎠
⎞
⎜⎜
⎝
⎛⎟⎟⎠
⎞⎜⎜⎝
⎛− ll
where λl =
lcrne MM
Mcrl = Critical elastic local beam buckling load including hole(s) Mne = defined in section above
Distortional Buckling The nominal flexural strength, Mnd, for distortional buckling shall be calculated in accordance with the following:
(a) For λdnet ≤0.673 Mnd = ynetM (b) For λdnet >0.673
Mnd = ynet
5.0
ynet
crd
5.0
ynet
crd MMM
MM22.01 ⎟
⎟⎠
⎞⎜⎜⎝
⎛⎟⎟
⎠
⎞
⎜⎜
⎝
⎛
⎟⎟⎠
⎞⎜⎜⎝
⎛−
where λdnet =
crdynet MM
Mcrd = Critical elastic distortional beam buckling load including hole(s)
388
Option 3: Cap Mnl and Mnd, otherwise no strength change, include hole(s) in Mcr This method puts bounds in place and assumes local-global interaction happens at full Mne
Lateral-Torsional Buckling The nominal flexural strength, Mne, for lateral-torsional buckling shall be calculated in accordance with the following:
(a) for Mcre < 0.56 My Mne = Mcre
(b) for 2.78My ≥ Mcre ≥ 0.56My ⎟⎟⎠
⎞⎜⎜⎝
⎛−=
cre
yyne M36
M101M
910M
(c) for Mcre >2.78My Mne= My
where Mcre= Critical elastic global beam buckling load … (including hole(s))
Local Buckling The nominal flexural strength, Mnl, for local buckling shall be calculated in accordance with the following:
(a) For λl ≤0.776 Mnl = Mne ≤Mynet
(b) For λl > 0.776
Mnl = ne
4.0
ne
cr
4.0
ne
cr MMM
MM
15.01 ⎟⎟⎠
⎞⎜⎜⎝
⎛⎟⎟
⎠
⎞
⎜⎜
⎝
⎛⎟⎟⎠
⎞⎜⎜⎝
⎛− ll
where λl =
lcrne MM
Mcrl = Critical elastic local beam buckling load including hole(s) Mne= defined in section above
Mynet = SfnetFy≥0.80My
Sfnet = Section modulus at the hole(s) referenced to the extreme fiber at first yield
Distortional Buckling The nominal flexural strength, Mnd, for distortional buckling shall be calculated in accordance with the following:
(a) For λd ≤0.673 Mnd = My≤ Mynet (b) For λd >0.673
Mnd = y
5.0
y
crd
5.0
y
crd MMM
MM
22.01 ⎟⎟⎠
⎞⎜⎜⎝
⎛⎟⎟
⎠
⎞
⎜⎜
⎝
⎛
⎟⎟⎠
⎞⎜⎜⎝
⎛−
where λd =
crdy MM
Mcrd = Critical elastic distortional beam buckling load including hole(s)
389
Option 4: Cap Mnl, transition Mnd, include hole(s) in Mcr determinations This method puts bounds and transition in place, assumes local-global interaction at full Mne
Lateral-Torsional Buckling The nominal flexural strength, Mne, for lateral-torsional buckling shall be calculated in accordance with the following:
(a) for Mcre < 0.56 My Mne = Mcre
(b) for 2.78My ≥ Mcre ≥ 0.56My ⎟⎟⎠
⎞⎜⎜⎝
⎛−=
cre
yyne M36
M101M
910M
(c) for Mcre >2.78My Mne= My
where Mcre= Critical elastic global beam buckling load … (including hole(s))
Local Buckling The nominal flexural strength, Mnl, for local buckling shall be calculated in accordance with the following:
(a) For λl ≤0.776 Mnl = Mne ≤Mynet
(b) For λl > 0.776
Mnl = ne
4.0
ne
cr
4.0
ne
cr MMM
MM
15.01 ⎟⎟⎠
⎞⎜⎜⎝
⎛⎟⎟
⎠
⎞
⎜⎜
⎝
⎛⎟⎟⎠
⎞⎜⎜⎝
⎛− ll
where λl =
lcrne MM
Mcrl = Critical elastic local beam buckling load including hole(s) Mne= defined in section above
Mynet = SfnetFy≥0.80My
Sfnet = Section modulus at the hole(s) referenced to the extreme fiber at first yield
Distortional Buckling The nominal flexural strength, Mnd, for distortional buckling shall be calculated in with the following:
(a) For λd ≤λd1 Mnd = ynetM (cap on column strength) (b) For λd1<λd≤λd2
Mnd = d
1d2d
2dynetynet
MMM λ
λλ ⎟⎟⎠
⎞⎜⎜⎝
⎛−
−− (yield control transition)
(c) For λd > λd2
Mnd = y
5.0
y
crd
5.0
y
crd MMM
MM
22.01 ⎟⎟⎠
⎞⎜⎜⎝
⎛⎟⎟
⎠
⎞
⎜⎜
⎝
⎛
⎟⎟⎠
⎞⎜⎜⎝
⎛− (existing DSM distortional curve)
where λd =
crdy MM
λd1 = )MM(673.0 yynet
λd2 = ( )( )7.0MM7.1673.0 7.1yynet −−
Μd2 = ( )( )( ) y2d2d M1122.01 λλ−
Mynet = SfnetFy≥0.80My (limit reduction of the net section to 0.8My)
Sfnet = Section modulus at the hole(s) referenced to the extreme fiber at first yield Mcrd = Critical elastic distortional beam buckling load including hole(s)
390
Option 5: Transition Mnl (Option A), transition Mnd, include hole(s) in Mcr determinations This method puts bounds and transition in place, assumes local-global interaction at full Mne
Lateral-Torsional Buckling The nominal flexural strength, Mne, for lateral-torsional buckling shall be calculated in accordance with the following:
(a) for Mcre < 0.56 My Mne = Mcre
(b) for 2.78My ≥ Mcre ≥ 0.56My ⎟⎟⎠
⎞⎜⎜⎝
⎛−=
cre
yyne M36
M101M
910M
(c) for Mcre >2.78My Mne= My
where Mcre= Critical elastic global beam buckling load … (including hole(s))
Local Buckling The nominal flexural strength, Mnl, for local buckling shall be calculated in accordance with the following:
(a) For λl ≤λl1 Mnl = Mne≤ Mynet (cap on beam strength)
(b) For λl1<λl≤λl2
Mnl = ( ) ⎟⎟⎠
⎞⎜⎜⎝
⎛−−
⎟⎟⎠
⎞⎜⎜⎝
⎛−−
12
122ynetynet MMM
ll
ll
l
ll λλ
λλλλ (nonlinear yield transition when Mynet/Mne ≤1)
(c) For λl > λl2
Mnl = ne
4.0
ne
cr
4.0
ne
cr MMM
MM
15.01 ⎟⎟⎠
⎞⎜⎜⎝
⎛⎟⎟
⎠
⎞
⎜⎜
⎝
⎛⎟⎟⎠
⎞⎜⎜⎝
⎛− ll (DSM local buckling curve, unchanged)
where λl =
lcrne MM
λl1 = )MM(776.0 neynet≤0.776
λl2 = ( )( )4.1MM4.2776.0 5.3neynet −− , Mynet/Mne≤1
= 0.776, Mynet/Mne>1 (no transition when Mynet/Mne >1 ) Μl2 = ( )( )( ) ne
8.02
8.02 M1115.01 ll λλ−
Mynet = SfnetFy≥0.80My (limit reduction of the net section to 0.8My)
Sfnet = Section modulus at the hole(s) referenced to the extreme fiber at first yield Mcrl = Critical elastic local beam buckling load including hole(s)
Distortional Buckling
Same as Option 4
391
Option 6: Transition Mnl (Option B), transition Mnd, include hole(s) in Mcr determinations This method puts bounds and transition in place, assumes local-global interaction at full Mne
Flexural, Torsional, or Torsional-Flexural Buckling The nominal axial strength, Pne, for flexural, … or torsional- flexural buckling is
for 5.1c ≤λ Pne = yP658.02c ⎟
⎠⎞⎜
⎝⎛ λ
for λc > 1.5 crey2c
ne P877.0P877.0P =⎟⎟⎠
⎞⎜⎜⎝
⎛
λ=
where λc = crey PP
Pcre= Critical elastic global column buckling load … (including hole(s)) Py = AgFy Ag = gross area of the column
Local Buckling The nominal axial strength, Mnl, for local buckling shall be calculated in accordance with the following:
(a) For λl ≤λl1 Mnl = Mynet (Mne /My) (cap on column strength)
(b) For λl1<λl≤λl2
Mnl = ⎟⎟⎠
⎞⎜⎜⎝
⎛−−
⎟⎟⎠
⎞⎜⎜⎝
⎛⎟⎟⎠
⎞⎜⎜⎝
⎛−⎟
⎟⎠
⎞⎜⎜⎝
⎛−⎟
⎟⎠
⎞⎜⎜⎝
⎛
12
122
y
neynet
y
neynet M
MMM
MMM
ll
ll
l
ll λλ
λλλλ (nonlinear yield transition)
(c) For λl > λl2
Mnl = ne
4.0
ne
cr
4.0
ne
cr MMM
MM
15.01 ⎟⎟⎠
⎞⎜⎜⎝
⎛⎟⎟
⎠
⎞
⎜⎜
⎝
⎛⎟⎟⎠
⎞⎜⎜⎝
⎛− ll (DSM local buckling curve, unchanged)
where λl =
lcrne MM
λl1 = )MM(776.0 yynet≤0.776
λl2 = ( )( )4.1MM4.2776.0 5.3yynet −−
Μl2 = ( )( )( ) ne8.0
28.0
2 M1115.01 ll λλ− Mynet = SfnetFy≥0.80My (limit reduction of the net section to 0.8My)
Sfnet = Section modulus at the hole(s) referenced to the extreme fiber at first yield Mcrl = Critical elastic local beam buckling load including hole(s)
Distortional Buckling
Same as Option 4
392
2.5 115B DSM comparison to beam test simulation database
The six DSM prediction options for cold‐formed steel beams with holes are
evaluated with the simulated laterally braced beam experiment database developed in
Section 1812H8.2.1 and summarized in 1813HAppendix L. The simulated data is compared against
DSM predictions while evaluating data trends against member slenderness and hole size
(Inet/Ig), and span to depth ratio (L/H).
1814HFigure 8.51 and 1815HFigure 8.52 compare the simulated test data to predictions for local
and distortional buckling‐controlled beam failures. Option 1 is identical to the existing
DSM approach for beams without holes, except the critical elastic buckling loads (Mcrl,
Mcrd, and Mcre) are determined with the influence of holes. Option 1 is observed to be a
accurate predictor of local buckling controlled failure strengths, although distortional
predictions are conservative when λd is high and unconservative by as much as 20% asλd
decreases below 1.5 (see 1816HFigure 8.52). The unconservative predictions occur because
Option 1 does not account for the column strength limit Mynet, nor does it account for a
transition from an elastic buckling controlled failure to a yield‐controlled failure at the
net section.
Option 2 is observed to be a conservative predictor in 1817HFigure 8.51 and 1818HFigure 8.52 for
high λl, and λd and demonstrates improved accuracy over Option 1 when slenderness
decreases and hole size increases (see 1819HFigure 8.52). Option 2 replaces Mynet everywhere
within the existing DSM formulation, which has the effect of increasing λl and λd and
decreasing predicted strength. Option 3 test‐to‐predicted trends are similar to Option 1
with increasingly unconservative predictions as slenderness decreases, demonstrating
393
that the Mynet limits on Mnl and Mnd in Option 3 are not fully effective at capturing the
yield transition to the net section. Option 4 is identical to Option 3 except the yield
transition on the DSM distortional curve is employed to provide a more accurate
prediction of the net‐section yielding influence. Option 4 demonstrates an improvement
in distortional buckling‐controlled prediction accuracy when λd < 1, although the
strength of 11 beams are underpredicted by up to 15% when λd=1.3. Option 5 accurately
predicts the strength of these 11 beams with the added transition on the local buckling
design curve. (Option 6 is the same as Option 5 because the beams considered are
laterally‐braced).
1820HTable 8.1 summarizes the test‐to‐predicted ratio statistics and strength reduction
factor φ for the six DSM options (see Eq. 1821H(8.1) for a definition of φ). No one option stands
out above the rest when studying the table, although the observations from 1822HFigure 8.51
and 1823HFigure 8.52 support Options 3,4, and 5(6) as the methods most closely tied to
underlying collapse mechanisms at ultimate limit state. The observations from this
comparison will be employed along with the DSM comparison to the beam experimental
database in the next section to support the recommended DSM modifications.
Table 8.4 DSM test‐to‐predicted statistics for laterally braced beam simulations Option Description
Mean SD φ # of tests Mean SD φ # of tests1 My everywhere 1.07 0.09 0.89 44 1.06 0.13 0.86 1602 Mynet everywhere 1.05 0.10 0.88 50 1.07 0.12 0.86 1543 Cap Mnl, Mnd 1.07 0.09 0.89 44 1.06 0.13 0.86 1604 Transition Mnd, Cap Mnl 1.07 0.09 0.89 44 1.06 0.13 0.86 160
5,6 Transition Mnd and Mnl (Option A, B) 1.01 0.11 0.87 72 1.09 0.12 0.87 132
Local buckling Distortional buckling
394
0 0.5 1 1.5 2 2.5 30
0.5
1
1.5Option 1 - My everywhere
λl=(My/Mcrl
)0.5
Mn/M
y
0 0.5 1 1.5 2 2.5 30
0.5
1
1.5Option 2 - Mynet everywhere
λl=(My/Mcrl
)0.5
Mn/M
y
0 0.5 1 1.5 2 2.5 30
0.5
1
1.5Option 3 - cap Mnl
, Mnd
λl=(My/Mcrl
)0.5
Mn/M
y
0 0.5 1 1.5 2 2.5 30
0.5
1
1.5Option 4 - cap Mnl
, transition Mnd
λl=(My/Mcrl
)0.5
Mn/M
y
DSM Local PredictionLocal Controlled Test
0 0.5 1 1.5 2 2.5 30
0.5
1
1.5Option 5 - transition Mnl
(Option A), transition Mnd
λl=(My/Mcrl
)0.5
Mn/M
y
0 0.5 1 1.5 2 2.5 30
0.5
1
1.5Option 6 - transition Mnl
(Option B), transition Mnd
λl=(My/Mcrl
)0.5
Mn/M
y
Figure 8.54 Comparison of simulated test strengths to predictions for laterally braced beams with local buckling‐controlled failures as a function of local
slenderness
395
0 0.5 1 1.5 2 2.5 30
0.5
1
1.5Option 1 - My everywhere
λd=(My/Mcrd)0.5
Mn/M
y
0 0.5 1 1.5 2 2.5 30
0.5
1
1.5Option 2 - Mynet everywhere
λdnet=(Mynet/Mcrd)0.5
Mn/M
y
DSM Dist. PredictionDist. Controlled Test
0 0.5 1 1.5 2 2.5 30
0.5
1
1.5Option 3 - cap Mnl
, Mnd
λd=(My/Mcrd)0.5
Mn/M
y
0 0.5 1 1.5 2 2.5 30
0.5
1
1.5Option 4 - cap Mnl
, transition Mnd
λd=(My/Mcrd)0.5
Mn/M
y
0 0.5 1 1.5 2 2.5 30
0.5
1
1.5Option 5 - transition Mnl
(Option A), transition Mnd
λd=(My/Mcrd)0.5
Mn/M
y
0 0.5 1 1.5 2 2.5 30
0.5
1
1.5Option 6 - transition Mnl
(Option B), transition Mnd
λd=(My/Mcrd)0.5
Mn/M
y
Figure 8.55 Comparison of simulated test strengths to predictions for laterally braced beams with distortional buckling‐controlled failures as a function of
distortional slenderness
396
0 0.5 1 1.5 2 2.5 30
0.5
1
1.5
Option 1 - My everywhere
λl=(My/Mcrl
)0.5
Mte
st/M
n
0 0.5 1 1.5 2 2.5 30
0.5
1
1.5
Option 2 - Mynet everywhere
λl=(My/Mcrl
)0.5
Mte
st/M
n
Local Controlled
0 0.5 1 1.5 2 2.5 30
0.5
1
1.5
Option 3 - cap Mnl, Mnd
λl=(My/Mcrl
)0.5
Mte
st/M
n
0 0.5 1 1.5 2 2.5 30
0.5
1
1.5
Option 4 - cap Mnl, transition Mnd
λl=(My/Mcrl
)0.5
Mte
st/M
n
0 0.5 1 1.5 2 2.5 30
0.5
1
1.5
Option 5 - transition Mnl (Option A), transition Mnd
λl=(My/Mcrl
)0.5
Mte
st/M
n
0 0.5 1 1.5 2 2.5 30
0.5
1
1.5
Option 6 - transition Mnl (Option B), transition Mnd
λl=(My/Mcrl
)0.5
Mte
st/M
n
Figure 8.56 Test‐to‐predicted ratios for local buckling‐controlled simulated laterally braced beam failures as a function of local slenderness
397
0 0.5 1 1.5 2 2.5 30
0.5
1
1.5
Option 1 - My everywhere
λd=(My/Mcrd)0.5
Mte
st/M
n
0 0.5 1 1.5 2 2.5 30
0.5
1
1.5
Option 2 - Mynet everywhere
λdnet=(Mynet/Mcrd)0.5
Mte
st/M
n
Distortional Controlled
0 0.5 1 1.5 2 2.5 30
0.5
1
1.5
Option 3 - cap Mnl, Mnd
λd=(My/Mcrd)0.5
Mte
st/M
n
0 0.5 1 1.5 2 2.5 30
0.5
1
1.5
Option 4 - cap Mnl, transition Mnd
λd=(My/Mcrd)0.5
Mte
st/M
n
0 0.5 1 1.5 2 2.5 30
0.5
1
1.5
Option 5 - transition Mnl (Option A), transition Mnd
λd=(My/Mcrd)0.5
Mte
st/M
n
0 0.5 1 1.5 2 2.5 30
0.5
1
1.5
Option 6 - transition Mnl (Option B), transition Mnd
λd=(My/Mcrd)0.5
Mte
st/M
n
Figure 8.57 Test‐to‐predicted ratios for distortional buckling‐controlled simulated laterally braced beam failures as a function of distortional slenderness
398
0.5 0.6 0.7 0.8 0.9 10
0.5
1
1.5
Option 1 - My everywhere
Inet/Ig
Mte
st/M
n
0.5 0.6 0.7 0.8 0.9 10
0.5
1
1.5
Option 2 - Mynet everywhere
Inet/Ig
Mte
st/M
n
Local Controlled
0.5 0.6 0.7 0.8 0.9 10
0.5
1
1.5
Option 3 - cap Mnl, Mnd
Inet/Ig
Mte
st/M
n
0.5 0.6 0.7 0.8 0.9 10
0.5
1
1.5
Option 4 - cap Mnl, transition Mnd
Inet/Ig
Mte
st/M
n
0.5 0.6 0.7 0.8 0.9 10
0.5
1
1.5
Option 5 - transition Mnl (Option A), transition Mnd
Inet/Ig
Mte
st/M
n
0.5 0.6 0.7 0.8 0.9 10
0.5
1
1.5
Option 6 - transition Mnl (Option B), transition Mnd
Inet/Ig
Mte
st/M
n
Figure 8.58 Test‐to‐predicted ratios for simulated local buckling‐controlled laterally braced beam failures as a function of net cross‐sectional moment of inertia to
gross cross‐sectional moment of inertia
399
0.5 0.6 0.7 0.8 0.9 10
0.5
1
1.5
Option 1 - My everywhere
Inet/Ig
Mte
st/M
n
0.5 0.6 0.7 0.8 0.9 10
0.5
1
1.5
Option 2 - Mynet everywhere
Inet/Ig
Mte
st/M
n
Distortional Controlled
0.5 0.6 0.7 0.8 0.9 10
0.5
1
1.5
Option 3 - cap Mnl, Mnd
Inet/Ig
Mte
st/M
n
0.5 0.6 0.7 0.8 0.9 10
0.5
1
1.5
Option 4 - cap Mnl, transition Mnd
Inet/Ig
Mte
st/M
n
0.5 0.6 0.7 0.8 0.9 10
0.5
1
1.5
Option 5 - transition Mnl (Option A), transition Mnd
Inet/Ig
Mte
st/M
n
0.5 0.6 0.7 0.8 0.9 10
0.5
1
1.5
Option 6 - transition Mnl (Option B), transition Mnd
Inet/Ig
Mte
st/M
n
Figure 8.59 Test‐to‐predicted ratios for simulated distortional buckling‐controlled laterally braced beam failures as a function of net cross‐sectional moment of
inertia to gross cross‐sectional moment of inertia
400
0 2 4 6 8 100
0.5
1
1.5
Option 1 - My everywhere
L/H
Mte
st/M
n
0 2 4 6 8 100
0.5
1
1.5
Option 2 - Mynet everywhere
L/H
Mte
st/M
n
Local Controlled
0 2 4 6 8 100
0.5
1
1.5
Option 3 - cap Mnl, Mnd
L/H
Mte
st/M
n
0 2 4 6 8 100
0.5
1
1.5
Option 4 - cap Mnl, transition Mnd
L/H
Mte
st/M
n
0 2 4 6 8 100
0.5
1
1.5
Option 5 - transition Mnl (Option A), transition Mnd
L/H
Mte
st/M
n
0 2 4 6 8 100
0.5
1
1.5
Option 6 - transition Mnl (Option B), transition Mnd
L/H
Mte
st/M
n
Figure 8.60 Test‐to‐predicted ratios for simulated local buckling‐controlled laterally braced beam failures as a function of column length, L, to beam depth, H
401
0 2 4 6 8 100
0.5
1
1.5
Option 1 - My everywhere
L/H
Mte
st/M
n
0 2 4 6 8 100
0.5
1
1.5
Option 2 - Mynet everywhere
L/H
Mte
st/M
n
Distortional Controlled
0 2 4 6 8 100
0.5
1
1.5
Option 3 - cap Mnl, Mnd
L/H
Mte
st/M
n
0 2 4 6 8 100
0.5
1
1.5
Option 4 - cap Mnl, transition Mnd
L/H
Mte
st/M
n
0 2 4 6 8 100
0.5
1
1.5
Option 5 - transition Mnl (Option A), transition Mnd
L/H
Mte
st/M
n
0 2 4 6 8 100
0.5
1
1.5
Option 6 - transition Mnl (Option B), transition Mnd
L/H
Mte
st/M
n
Figure 8.61 Test‐to‐predicted ratios for simulated distortional buckling‐controlled laterally braced beam failures as a function of column length, L, to H
402
2.6 116BDSM comparison to experimental beam database
The six DSM options are now compared to the laterally braced beam experiment
database first assembled in Section 1824H4.3.1. The database contains the elastic buckling
properties of each beam, including the presence of holes and the influence of boundary
conditions, as well as the tested strengths. 1825HFigure 8.62 through 1826HFigure 8.65 compare the
experiment strengths to DSM predictions for local and distortional buckling‐controlled
beam failures. (The local and distortional slenderness is obtained with the “pure” local
and distortional elastic buckling loads L and D in this study, not the LH and DH modes
described in Section 1827H4.3). The tested strengths are lower than the predictions over a wide
range of local and distortional slenderness. These trends were first observed in a
preliminary DSM comparison (Moen and Schafer 2007a), and possible reasons for the
difference between test and predictions were hypothesized, including experimental
error, error in the determination of elastic buckling loads, and the influence of the angle
straps on the calculation of the distortional critical elastic buckling load. The beams in
the database have relatively small holes, with Inet/Ig ranging from 0.96 to 0.99 as shown in
1828HFigure 8.64 and 1829HFigure 8.65, which suggests that the presence of holes should not have a
significant impact on tested strength. The test‐to‐predicted statistics are the same for the
six DSM options as shown in 1830HTable 8.5. It is concluded that the experimental database,
in its current form, cannot be used to evaluate the proposed DSM modifications. Future
work is planned to investigate the differences between the DSM predictions and tested
strengths for this data. In addition, more recent tests on cold‐formed steel beams with
403
holes will be added to the database. Experiments on beams with larger holes are also
needed.
Table 8.5 DSM test‐to‐predicted ratio statistics for beam experiments Option Description
Mean SD φ # of tests Mean SD φ # of tests1 My everywhere 0.88 0.12 0.85 55 0.87 0.14 0.81 892 Mynet everywhere 0.88 0.12 0.85 55 0.87 0.14 0.81 893 Cap Mnl, Mnd 0.88 0.12 0.85 55 0.87 0.14 0.81 894 Transition Mnd, Cap Mnl 0.88 0.12 0.85 55 0.87 0.14 0.81 89
5,6 Transition Mnd and Mnl (Option A, B) 0.88 0.12 0.85 55 0.87 0.14 0.81 89
Local buckling Distortional buckling
404
0 0.5 1 1.5 2 2.5 30
0.5
1
1.5Option 1 - My everywhere
λl=(My/Mcrl
)0.5
Mn/M
y
0 0.5 1 1.5 2 2.5 30
0.5
1
1.5Option 2 - Mynet everywhere
λl=(My/Mcrl
)0.5
Mn/M
y
0 0.5 1 1.5 2 2.5 30
0.5
1
1.5Option 3 - cap Mnl
, Mnd
λl=(My/Mcrl
)0.5
Mn/M
y
0 0.5 1 1.5 2 2.5 30
0.5
1
1.5Option 4 - cap Mnl
, transition Mnd
λl=(My/Mcrl
)0.5
Mn/M
y
DSM Local PredictionLocal Controlled Test
0 0.5 1 1.5 2 2.5 30
0.5
1
1.5Option 5 - transition Mnl
(Option A), transition Mnd
λl=(My/Mcrl
)0.5
Mn/M
y
0 0.5 1 1.5 2 2.5 30
0.5
1
1.5Option 6 - transition Mnl
(Option B), transition Mnd
λl=(My/Mcrl
)0.5
Mn/M
y
Figure 8.62 Comparison of experimental test strengths to predictions for laterally braced beams with local buckling‐controlled failures as a function of local
slenderness
405
0 0.5 1 1.5 2 2.5 30
0.5
1
1.5Option 1 - My everywhere
λd=(My/Mcrd)0.5
Mn/M
y
0 0.5 1 1.5 2 2.5 30
0.5
1
1.5Option 2 - Mynet everywhere
λdnet=(Mynet/Mcrd)0.5
Mn/M
y
DSM Dist. PredictionDist. Controlled Test
0 0.5 1 1.5 2 2.5 30
0.5
1
1.5Option 3 - cap Mnl
, Mnd
λd=(My/Mcrd)0.5
Mn/M
y
0 0.5 1 1.5 2 2.5 30
0.5
1
1.5Option 4 - cap Mnl
, transition Mnd
λd=(My/Mcrd)0.5
Mn/M
y
0 0.5 1 1.5 2 2.5 30
0.5
1
1.5Option 5 - transition Mnl
(Option A), transition Mnd
λd=(My/Mcrd)0.5
Mn/M
y
0 0.5 1 1.5 2 2.5 30
0.5
1
1.5Option 6 - transition Mnl
(Option B), transition Mnd
λd=(My/Mcrd)0.5
Mn/M
y
Figure 8.63 Comparison of experimental test strengths to predictions for laterally braced beams with distortional buckling‐controlled failures as a function of
distortional slenderness
406
0.5 0.6 0.7 0.8 0.9 10
0.5
1
1.5
Option 1 - My everywhere
Inet/Ig
Mte
st/M
n
0.5 0.6 0.7 0.8 0.9 10
0.5
1
1.5
Option 2 - Mynet everywhere
Inet/Ig
Mte
st/M
n
Local Controlled
0.5 0.6 0.7 0.8 0.9 10
0.5
1
1.5
Option 3 - cap Mnl, Mnd
Inet/Ig
Mte
st/M
n
0.5 0.6 0.7 0.8 0.9 10
0.5
1
1.5
Option 4 - cap Mnl, transition Mnd
Inet/Ig
Mte
st/M
n
0.5 0.6 0.7 0.8 0.9 10
0.5
1
1.5
Option 5 - transition Mnl (Option A), transition Mnd
Inet/Ig
Mte
st/M
n
0.5 0.6 0.7 0.8 0.9 10
0.5
1
1.5
Option 6 - transition Mnl (Option B), transition Mnd
Inet/Ig
Mte
st/M
n
Figure 8.64 Test‐to‐predicted ratios for experimental local buckling‐controlled laterally braced beam failures as a function of net cross‐sectional moment of inertia
to gross cross‐sectional moment of inertia
407
0.5 0.6 0.7 0.8 0.9 10
0.5
1
1.5
Option 1 - My everywhere
Inet/Ig
Mte
st/M
n
0.5 0.6 0.7 0.8 0.9 10
0.5
1
1.5
Option 2 - Mynet everywhere
Inet/Ig
Mte
st/M
n
Distortional Controlled
0.5 0.6 0.7 0.8 0.9 10
0.5
1
1.5
Option 3 - cap Mnl, Mnd
Inet/Ig
Mte
st/M
n
0.5 0.6 0.7 0.8 0.9 10
0.5
1
1.5
Option 4 - cap Mnl, transition Mnd
Inet/Ig
Mte
st/M
n
0.5 0.6 0.7 0.8 0.9 10
0.5
1
1.5
Option 5 - transition Mnl (Option A), transition Mnd
Inet/Ig
Mte
st/M
n
0.5 0.6 0.7 0.8 0.9 10
0.5
1
1.5
Option 6 - transition Mnl (Option B), transition Mnd
Inet/Ig
Mte
st/M
n
Figure 8.65 Test‐to‐predicted ratios for experimental distortional buckling‐controlled laterally braced beam failures as a function of net cross‐sectional moment of
inertia to gross cross‐sectional moment of inertia
408
2.7 117BRecommendations – DSM for beams with holes
Options 3, 4, 5, and 6 are presented as viable proposals for extending DSM to beams
with holes. This recommendation is based on the test‐to‐predicted statistics and data
trends from the simulation studies presented in Section 1831H8.2.5, and also considers the
effort to implement the modifications and their ease of use by design engineers. Option
3 accounts for the reduction in beam strength from the presence of holes by limiting Mnl
and Mnd to Mynet. This is a simple modification to implement in the Specification and
avoids additional calculation work for a design engineer (except for that required to
calculate the critical elastic buckling loads including the influence of the holes). Options
4 and 5 are refinements of Option 3, where the cap on Mnl and Mnd becomes a transition
from an elastic buckling controlled failure mode to a yield controlled failure at Mynet.
These two methods require additional effort from the designer when compared to
Option 3, but they have an important advantage. Options 4, 5, and 6 are more closely
tied to the failure mechanisms influencing column strength because they capture the
yield transition to the net section in their predictions. The transitions increase the
probability that strength will be accurately predicted for general beam and hole
geometries. Option 5 has the additional advantage of capturing the influence of a yield
transition for closed cross‐sections that do not experience distortional buckling.
Additional nonlinear finite element simulations and experiments are needed to validate
the proposed modifications to the Direct Strength Method for beams subject to lateral‐
torsional buckling at ultimate limit state.
409
Chapter 9
8BConclusions and proposed future work
9.1 50BConclusions
Proposed Direct Strength Method design equations are now in place for cold‐formed
steel members with holes. The development of the method was initiated with thin shell
finite element eigenbuckling studies in ABAQUS on thin plates and full cold‐formed
steel members with holes. The buckling of the unstiffened strips adjacent to a hole in a
thin plate influenced, and sometimes controlled, the critical elastic buckling stress of
individual cross‐section elements. Unstiffened strip buckling was also closely associated
with distortional buckling modes at the location of the holes in C‐section columns and
beams. Large holes and closely‐spaced holes locally stiffened thin rectangular plates
and the webs of C‐section columns, resulting in buckling away from the holes. The
elastic buckling studies led to useful design guidelines and tools, including hole spacing
limits (which prevent cumulative reductions in elastic stiffness along the length of a
410
member) and simplified elastic buckling prediction methods for local, distortional, and
global buckling developed as an alternative to finite element eigenbuckling analysis.
The viability of the DSM framework for cold‐formed steel members with holes was
established early in this research using existing test results and the elastic buckling
properties of cold‐formed steel column and beam specimens with holes. Additional
experimental work evaluated the influence of holes on the load‐deformation response
and failure mechanisms for short and intermediate length C‐section columns. During
the experiments, holes were observed to locally stiffen the web of the intermediate
length C‐section columns and prevented dynamic mode switching (from local buckling
to distortional buckling) near peak load. Holes were also observed to decrease post‐
peak ductility for columns when the hole size was large relative to the web width (e.g.,
the 362S162‐33 specimens).
Results from the experimental program were used to validate a nonlinear finite
element modeling protocol. A concerted effort was made to simulate the initial state of a
cold‐formed steel member in the protocol, including imperfection magnitudes based on
measurement statistics and residual stresses and initial plastic strains from the cold‐
forming process predicted with a mechanics‐based approach. The nonlinear finite
element modeling capability was used to construct a large database of simulated column
and beam experiments with a wide range of hole sizes, spacings, and C‐section
dimensions. Simulation results demonstrated that as cross‐section distortional or local
slenderness decreased, the failure of a cold‐formed steel member with holes occurred by
yielding and collapse of the unstiffened strips at the net cross‐section. Collapse of the
411
unstiffened strips sometimes triggered unstable global failure modes in columns with
large holes, i.e., as hole size approached Anet=0.60Ay. (Global instabilities caused by
yielding at peak load were not studied for beams with holes in this thesis, only laterally
braced beams were considered.) Modifications to the local and distortional DSM curves
were made to account for this unique net‐section failure mechanism with a deliberate
transition and cap on member strength. The final proposed DSM method for members
with holes was validated with existing experimental data and the simulated experiments
database.
9.2 51BFuture work Several interesting future research topics resulted from the elastic buckling studies,
experiments, and nonlinear finite element simulations in this thesis. Future research is
planned to follow up on many of these ideas and questions. The major points of future
study, organized by research topic, are listed below.
Thin‐shell finite element modeling in ABAQUS (Chapter 2)
The S9R5 meshing guidelines developed in this thesis were developed primary for eigenbuckling analyses. Meshing guidelines which ensure accurate results in nonlinear finite element simulations are also needed. Studies are ongoing to develop rules for determining the minimum number of through‐thickness finite element integration points, the mesh density required for linear and quadratic finite element formulations, and limits on initial element distortion and curvature.
412
Elastic buckling of cold‐formed steel cross‐sectional elements with holes (Chapter 3)
1. The simplified elastic buckling prediction method presented in this thesis for unstiffened elements loaded with uniaxial compression is empirically derived. A mechanics‐based unstiffened element prediction method is warranted as a topic of future research to improve the generality of the method.
2. An element‐based elastic buckling prediction method which accounts for stress gradients on unstiffened elements with holes is needed to address a design case engineers may encounter in practice.
3. Elastic buckling studies are planned to develop element‐based simplified methods for hole patterns found in storage racks.
4. The element‐based elastic buckling prediction methods provide a convenient method to calculate Fcr (the critical elastic buckling stress) for general hole shapes, sizes, and spacings for use in the AISI‐S100‐07 effective width method. Work is planned to evaluate introduce these simplified approaches into the effective width method.
Elastic buckling of cold‐formed steel members with holes (Chapter 4)
1. Yu and Davis, Ortiz‐Colberg, Rhodes and MacDonald, Rhodes and Schnieder, and Pu et al. performed tests on column specimens with multiple discrete holes or hole patterns. The elastic buckling properties and tested strengths of these specimens will be added to the experiment database, in addition to tests on rack sections.
2. Automated elastic buckling modal identification tools are needed to identify local, distortional, and global buckling modes in thin‐shell finite element eigenbuckling analysis. Research is ongoing to develop this capability with an implementation similar to that of the constrained finite strip method.
3. Work continues on the development and validation of the CUFSM elastic buckling approximate methods developed and the extension of these methods to members with hole patterns (e.g., storage racks). A general procedure for implementing CUFSM constraints in the local buckling prediction method is needed. Also, the current assumption that the warping torsion constant Cw=0 at a hole produces conservative global elastic buckling predictions for columns and beams. Additional research is needed to derive a mechanics‐based approximation for Cw at a hole.
Experiments on cold‐formed steel columns with holes (Chapter 5)
1. A more definitive method of measuring the base metal thickness of cold‐formed steel members with a zinc galvanic coating is needed. Current standard practice is to remove the zinc coating with hydrochloric acid or a ferric chloride solution. It is difficult to know when all of the zinc has been removed though since the
413
zinc chemically interacts with the base metal during the initial application. Experiments are planned to determine the influence of the zinc coating on ultimate strength.
2. Research work is planned to evaluate the influence of sheet coiling on the measured yield stress in tensile coupons. It has been hypothesized by Professor Rasmussen at the University of Sydney that the same coiling curvature which causes residual stresses in cold‐formed steel members also affects yield stress measurements in tensile tests.
Residual stresses and plastic strains in cold‐formed steel members (Chapter 6)
1. Experimental work is planned to validate the prediction model presented in Chapter 6 relating coiling residual stresses to the coiling radius, sheet thickness, and yield stress.
2. Research is ongoing to evaluate how ABAQUS metal plasticity laws use the residual stress and initial plastic strain information and to determine if kinematic hardening or a different mixed hardening rule is required to accurately simulate the cold‐work of forming effect on load‐deformation response.
3. Nonlinear finite element studies are planned to identify the influence of through‐thickness residual stresses and plastic strains on the load‐deformation response, ultimate strength, and failure mechanisms of cold‐formed steel beams and columns.
4. Hancock et al. provides a method which accounts for the cold‐work of forming in the corners of cold‐formed steel cross‐sections when calculating ultimate strength (Hancock et al. 2001). The research in Chapter 6 provides new insight into the relationship between residual stresses and initial plastic strains from the manufacturing process. Research work is planned to revisit Hancock’s cold‐work of forming method to determine if it can be supplemented with this new research.
5. The current residual stress prediction method assumes an elastic‐perfectly plastic material model. Research work is ongoing to introduce the effect of steel strain hardening into the prediction method.
Nonlinear finite element modeling of cold‐formed steel members (Chapter 7)
1. The use of measured imperfection magnitudes instead of statistical distributions is warranted as a topic of future study, especially the use of a flared cross‐section, including flange‐web angles off of 90 degrees.
2. Initiating plasticity in ABAQUS at the material’s proportional limit reduced the predicted strength by up to 20% when compared to experiments in Chapter 7. A study is planned to simulate a single finite element in tension to evaluate the ABAQUS implementation of metal plasticity and determine the source of the discrepancy.
414
The Direct Strength Method for members with holes (Chapter 8)
1. Additional validation studies are planned to compare the proposed DSM Holes methodology to the AISI‐S100‐07 effective width design method.
2. Nonlinear finite element studies of other DSM prequalified cross‐sections (e.g., Z‐sections and hat sections) as well as rack sections with hole patterns are also planned to expand the simulation database.
415
References
ABAQUS. (2007a). ʺABAQUS/Standard Version 6.7‐1.ʺ, Dassault Systèmes,
435Hhttp://www.simulia.com/, Providence, RI. ABAQUS. (2007b). ʺPersonal communication, S9R5 element section numbers.ʺ Dassault
Systèmes, 436Hhttp://www.simulia.com/, Providence, RI. AISI‐S100. (2007). North American Specification for the Design of Cold‐Formed Steel Structural
Members, American Iron and Steel Institute, Washington, D.C. AISI‐TS‐2‐02. (2001). ʺStub Column Test Method for Effective Area of Cold‐formed Steel
Columns.ʺ 2001 North American Cold Formed Steel Specification, American Iron and Steel Institute, Washington, D.C.
AISI. (2006). Direct Strength Method Design Guide, American Iron and Steel Institute,
Washington, D.C. ASTM. (2004). ʺE 8M‐04, Standard Test Methods for Tension Testing of Metallic Materials
(Metric).ʺ ASTM International, West Conshohocken, PA. Batson, K.D. (1992). ʺFlexural Behavior of Webs with Openings,ʺ M.S. Thesis, University of
Missouri‐Rolla, Rolla. Bernard, E.S. (1993). ʺFlexural Behaviour of Cold‐Formed Profiled Steel Decking,ʺ Ph.D. Thesis,
University of Sydney, Australia, Sydney. Chajes, A. (1974). Principles of Structural Stability, Prentice Hall College Div, Englewood Cliffs,
NJ. Chen, W.F., and Han, D.J. (1988). Plasticity for Structural Engineers, Springer‐Verlag, New York,
NY. Cook, R.D. (1989). Concepts and Applications of Finite Element Analysis, J. Wiley & Sons, New
York, NY. CRC. (2003). Standard Mathematical Tables and Formulae, CRC Press, New York, NY. Crisfield, M.A. (1981). ʺA fast incremental\iteration solution procedure that handles snap‐
through.ʺ Computers and Structures, 13, 55‐62. Dat, D.T. (1980). ʺThe Strength of Cold‐Formed Steel Columns.ʺ Cornell University Department
of Structural Engineering Report No. 80‐4, Ithaca, NY.
416
De Batista, E.M., and Rodrigues, F.C. (1992). ʺResidual stress measurements on cold‐formed profiles.ʺ Experimental Techniques, 16(5), 25‐29.
Galambos, T. (1998a). ʺAppendix B.3, Technical Memorandum No.3: Stub Column Test
Procedure.ʺ Guide to Stability Design Criteria for Metal Structures, 5th Edition, John Wiley & Sons, New York, NY.
Galambos, T. (1998b). Guide to Stability Design Criteria for Metal Structures, 5th Edition, John
Wiley & Sons, New York, NY. Hancock, G.J., Murray, T., and Ellifritt, D. (2001). Cold‐formed steel structures to the AISI
specification, Marcel Dekker, Inc., New York, NY. Hill, R. (1950). The Mathematical Theory of Plasticity, Oxford University Press, London, England. Ingvarsson, L. (1975). ʺCold‐forming residual stresses, effect of buckling.ʺ Proceedings, Third
Annual Specialty Conference on Cold‐Formed Steel Structures, University of Missouri‐Rolla, 85‐119.
Key, P., and Hancock, G.J. (1993). ʺA Theoretical Investigation of the Column Behaviour of
Cold‐Formed Square Hollow Sections.ʺ Thin‐Walled Structures, 16, 31‐64. Kwon, Y.B. (1992). ʺPost‐Buckling Behaviour of Thin‐Walled Sections,ʺ Ph.D. Thesis, University
of Sydney, Australia. Labview. (2005). ʺLabview, Version 8.ʺ National Instruments, 437Hwww.labview.com, Austin, TX. Mathworks. (2007). ʺMatlab 7.5.0 (R2007b).ʺ Mathworks, Inc., 438Hwww.mathworks.com. Moen, C.D., and Schafer, B.W. (2007a). ʺDirect strength design for cold‐formed steel members
with perforations, Progress Report #3.ʺ American Iron and Steel Institute, Washington, D.C.
Moen, C.D., and Schafer, B.W. (2007b). ʺDirect strength design for cold‐formed steel members
with perforations, Progress Report #4.ʺ American Iron and Steel Institute, Washington, D.C.
Oden, J.T., and Martins, J.A.C. (1985). ʺModels and computational methods for dynamic friction
phenomena.ʺ Computer Methods in Applied Mechanics and Engineering, 52, 527‐634. Ortiz‐Colberg, R.A. (1981). ʺThe Load Carrying Capacity of Perforated Cold‐Formed Steel
Columns,ʺ M.S. Thesis, Cornell University, Ithaca, NY. Porter, F.C. (1991). Zinc Handbook, Marcel Dekker, New York, NY. Powell, G., and Simons, J. (1981). ʺImproved iterative strategy for nonlinear structures.ʺ
International Journal for Numerical Methods in Engineering, 17, 1455‐1467.
417
Quach, W.M., Teng, J.G., and Chung, K.F. (2004 ). ʺResidual stresses in steel sheets due to coiling and uncoiling: a closed‐form analytical solution.ʺ Engineering Structures, 26, 1249‐1259.
Quach, W.M., Teng, J.G., and Chung, K.F. (2006 ). ʺFinite element predictions of residual
stresses in press‐braked thin‐walled steel sections.ʺ Engineering Structures, 28, 1609‐1619. Ramm, E. (1981). ʺStrategies for tracing nonlinear response near limit points.ʺ Nonlinear Finite
Element Analysis in Structural Mechanics: Proceedings of the Europe–US Workshop, 63‐89. Schafer, B.W. (1997). ʺCold‐formed steel behavior and design: analytical and numerical
modeling of elements and members with longitudinal stiffeners,ʺ Ph.D. Thesis, Cornell University, Ithaca, NY.
Schafer, B.W., and Ádàny, S. (2006). ʺBuckling analysis of cold‐formed steel members using
CUFSM: conventional and constrained finite strip methods.ʺ Eighteenth International Specialty Conference on Cold‐Formed Steel Structures, Orlando, FL.
Schafer, B.W., and Peköz, T. (1998). ʺComputational modeling of cold‐formed steel:
characterizing geometric imperfections and residual stresses.ʺ Journal of Constructional Research, 47, 193‐210.
Schafer, B.W., Sarawit, A., and Peköz, T. (2006). ʺComplex edge stiffeners for thin‐walled
members.ʺ Journal of Structural Engineering, 132(2), 212‐226. Schuster, R.M. (1992). ʺTesting of Perforated C‐Stud Sections in Bending.ʺ University of
Waterloo, Ontario, Canada. Shan, M.Y., and LaBoube, R.A. (1994). ʺBehavior of Web Elements With Openings Subjected to
Bending, Shear, and the Combination of Bending and Shear.ʺ Civil Engineering Study 94‐2, Cold‐formed Steel Series, University of Missouri‐Rolla, Rolla.
Shanley, F.R. (1957). Strength of Materials, McGraw‐Hill Book Company, New York, NY. Silvestre, N., and Camotim, D. (2005). ʺLocal‐plate and distortional post‐buckling behavior of
cold‐formed steel lipped channel columns with intermediate stiffeners.ʺ Seventeenth International Specialty Conference on Cold‐Formed Steel Structures: Recent Research and Developments in Cold‐Formed Steel Design and Construction, Orlando, FL, United States, 1‐18.
SSMA. (2001). ʺProduct Technical Information, ICBO ER‐4943P.ʺ Steel Stud Manufacturers
Association, 439Hwww.ssma.com. Timoshenko, S.P., Gere, James M. (1961). Theory of Elastic Stability, McGraw‐Hill, New York. US Steel. (1985). The Making, Shaping, and Treating of Steel, 10th Edition, Herbick & Held,
Pittsburgh, PA.
418
von Karman, T., Sechler, E.E., and Donnell, L.H. (1932). ʺThe strength of thin plates in compression.ʺ Transactions ASME, 54(APM 54‐5).
Wang, X.P., Lam, S.S.E., and Chung, K.F. (2006). ʺCross section distortion due to cutting of cold‐
formed steel lipped C‐section.ʺ Thin‐Walled Structures, 44, 271‐280. Weng, C.C., and Peköz, T. (1990). ʺResidual Stresses in Cold‐Formed Steel Members.ʺ ASCE
Journal of Structural Engineering, 116(6), 1611‐1625. Western States Clay Products Association. (2004). Design Guide for Anchored Brick Veneer over
Steel Stud Systems, 440Hwww.brick_wscpa.org, Seattle, WA. Winter, G. (1947). ʺStrength of thin steel compression flanges.ʺ Cornell University Engineering
Experiment Station Reprint No. 32. Yu, C. (2005). ʺDistortional buckling of cold‐formed steel members in bending,ʺ Ph.D. Thesis,
Johns Hopkins University, Baltimore. Yu, C., and Schafer, B.W. (2006). ʺDistortional buckling tests on cold‐formed steel beams.ʺ ASCE
Journal of Structural Engineering, 132(4), 515‐528. Yu, C., and Schafer, B.W. (2007). ʺEffect of Longitudinal Stress Gradients on Elastic Buckling of
Thin Plates.ʺ ASCE Journal of Structural Engineering, 133(4), 452‐463. Yu, W.W. (2000). Cold‐Formed Steel Design, John W. Wiley & Sons, New York, NY. Ziemian, R., and McGuire, W. (2005). ʺMASTAN version 3.0.4, 441Hwww.mastan.com.ʺ Bucknell
University, Lewisburg, PA.
419
Appendix A 9BABAQUS input file generator in Matlab
The finite element models in this thesis were generated with a custom Matlab
program which assembles a column or beam with any general cross‐section (input in
CUFSM‐style format) using nine node S9R5 thin‐shell finite elements. The user has the
ability to add holes at specific locations in the member, dictate the boundary conditions
and application of load, specify the material properties, and impose imperfection,
residual stresses and plastic strains to define a member’s initial state. Input files for
eigenbuckling analysis and nonlinear finite element simulations can be generated. The
program was used throughout this research to generate groups of ABAQUS input files
for parameter studies. The program setup used to generate the nonlinear finite element
models of the column experiments is provided here.
420
clear all close all sourceloc='C:\Documents and Settings\Cris\Desktop\cmoen\Cold Formed Steel - Holes Research\Fall 2007\runbuck development\Rev_6NL\jhab' %This example generates ABAQUS input files for the nonlinear %simulation of 12 different columns with SSMA C-sections and evenly spaced %slotted holes. The column boundary conditions are pinned-pinned warping free and the column is %loaded from both ends with a uniform compressive stress simulated as %consistent nodal loads on the first two sets of cross-section nodes. %25%CDF and 75% CDF imperfections are imposed on the member geometry with %CUFSM local and distortional buckling mode shapes. addpath([sourceloc '\functions\filewriting\']) addpath([sourceloc '\functions\holes\']) addpath([sourceloc '\functions\']) addpath([sourceloc '\templates\']) addpath([sourceloc '\']) load SSMAxsections %SSMA cross section info load SSMAnames %SSMA name list load SSMA_wvlengths %SSMA cFSM wavelengths for Pcrl, Pcrd load Ag %SSMA gross cross-sectional area %define the SSMA sections to create models for sections=[12 86 11 73 39 95 72 56 47 75 66 87] %define the imperfection magnitudes imptypes=[25 75] %define hole length (slotted holes considered here) Lhole=4 %define rough hole spacing, will be adjusted in holes section of file S=12 %define member lengths Lc=[34 88 24 74
421
42 78 66 56 32 74 40 80] %define hole depth such that Anet=0.70Ag Anetfactor=0.7 count=1 for i=1:length(sections) section_num=sections(i) for j=1:length(imptypes) %MEMBER LENGTH L=Lc(i) %MESH ALONG LENGTH nele=L*2; %NUMBER OF SECTION POINTS THROUGH THE THICKNESS sectionpoints=5 %CROSS-SECTION DIMENSIONS % Z % % A % X % D2 / I \ D1 % RT2/_S2___ S ___S1_\RT1 % \ | / % B2 \ | / B1 % \ | / % ___F2_\__________________/_F1___ ABAQUS Y AXIS % RB2 H RB1 %Dimensions are out-to-out, angles are in degrees, t is base metal + %coating thickness, tbare is base metal thickness % [H B1 B2 D1 D2 F1 F2 S1 S2 RB1 RB2 RT1 RT2 t tbare] dims=SSMAxsections(section_num,2:16) %calculate hole depth hhole=Ag(section_num)*(1-Anetfactor)./dims(15) %CROSS-SECTION MESHING %number of elements around the cross section %[D1 RT1 B1 RB1 H RB2 B2 RT2 D2] n=[2 2 2 2 16 2 2 2 2]; %CorZ=1 C-section, CorZ=2 Z-section CorZ=1 [node,elem]=cztemplate(CorZ,dims,n) nnodes=length(node(:,1)); %Number of FSM cross-section nodes %Determine FE number of nodes and increment nL=2*nele+1; %Number of FE nodes along the length %Determine the node numbering increment along the length if nnodes<100 FEsection_increment=100; %so along the length the numbering goes up by 100's else FEsection_increment=nnodes+1; end %ADD ADDITIONAL NODES nodeadd=[] %MATERIAL PROPERTIES
422
%steel matprops(1).name='MAT100'; matprops(1).elastic=[29500 0.3]; matprops(1).plastic=[58.6, 0 64.1517, 0.00342827 68.2188, 0.00842827 72.0304, 0.0134283 77.9752, 0.0234283 82.2224, 0.0334283 85.7249, 0.0434283 88.4053, 0.0534283 90.7405, 0.0634283 92.652, 0.0734283 94.3657, 0.0834283 95.8299, 0.0934283 97.2001, 0.103428 ]; %IMPERFECTIONS %*****IMPERFECTIONS***** %type=0 no imperfections %type=1 use mode shapes from ABAQUS results file %type=2 input from file %type 3 impose CUFSM shapes as imperfections imperfections.type=3; imperfections.filename=[]; imperfections.step=[]; imperfections.mode=[] t=dims(15) if imptypes(j)==25 if L>24 imperfections.magnitude=[0.14*t 0.64*t L/2000] imperfections.wavelength=[SSMA_wvlengths(section_num,1) SSMA_wvlengths(section_num,2) L] else imperfections.magnitude=[0.14*t 0.64*t] imperfections.wavelength=[SSMA_wvlengths(section_num,1) SSMA_wvlengths(section_num,2)] end elseif imptypes(j)==75 if L>24 imperfections.magnitude=[0.66*t 1.55*t L/1000] imperfections.wavelength=[SSMA_wvlengths(section_num,1) SSMA_wvlengths(section_num,2) L] else imperfections.magnitude=[0.66*t 1.55*t] imperfections.wavelength=[SSMA_wvlengths(section_num,1) SSMA_wvlengths(section_num,2)] end end imperfections.plumb=[] imperfections.member=[1] %1 for column, 2 for beam %DEFINE HOLES %Add holes to your member. %hole.type=1 circular %hole.type=2 rectangular %hole.type=3 slotted w\radial ends %hole.dimension=['width or length (ABAQUS x direction)' 'height or diameter'] %hole.location=['CUFSM cross section node (must be odd!)' 'longitudinal location' 'shift hole in direction of height'] %hole.thickness = thickness of finite elements making up hole, usually the same as the rest of the member %I've defined two slotted holes here in the web of the cross-section. %number of holes nhole=floor(L/S) if nhole<1
423
nhole=1 end %final hole spacing Sfinal=floor(L/nhole) spacing=Sfinal/2:Sfinal:L-Sfinal/2 hole.type=[3*ones(nhole,1)]; %define dimensions for slotted hole hole.dimension=[Lhole*ones(nhole,1) hhole*ones(nhole,1)]; %define location of hole in cross-section hole.location = [(length(node)+1)/2*ones(nhole,1) spacing' zeros(nhole,1)] hole.thickness = [dims(1,15)*ones(length(hole.type),1)] hole.material=[100*ones(length(hole.type),1)]; hole.groups=[100000+[1:length(hole.type)]]; hole.fill=[zeros(length(hole.type),1)]; %If you don't want holes, replace above with %hole=[ ] %MEMBER END LOADINGS %Loading notation is similar to CUFSM. Apply P for compression, M for %moment, or a combination of both. Compression at both ends of %column are %shown here. Loads are applied as consistent nodal loads in ABAQUS. end1load.P=1; end1load.Mxx=0; end1load.Mzz=0; end1load.M11=0; end1load.M22=0; end2load.P=-1; end2load.Mxx=0; end2load.Mzz=0; end2load.M11=0; end2load.M22=0; %CALCULATE CONSISTENT NODAL LOADS ON MEMBER ENDS***** unsymm=0 [end1cload, end2cload, A, Ixx]=consist_endloads(node,elem,end1load,end2load,unsymm, nL, FEsection_increment); %ABAQUS NODE SETS %Define these node sets to apply boundary conditions in ABAQUS % nodesetinfo={'nodeset name' [xlim1 xlim2 xint] [ylim1 ylim2 yint] [zlim1 zlim2 zint] exclude} %where nodes are grouped based on xlim1<=x<=xlim2 and ylim1<=y<=ylim2 and zlim1<=z<=zlim2. %Instead of ranges, assign xint,yint,zint to something other than zero to group nodes at specific x,y,and z %distance intervals % xlim1:xint:xlim2, ylim1:yint:ylim2, ylim1:yint:ylim2. %The exclude command can be used to exclude previously defined node sets from the current node set. %exclude = 0 all nodes in range are included in nodeset %exclude = m excludes nodeset m from current nodeset nodesetinfo={'ENDXZERO' [0 0 0] [-1000 1000 0] [-1000 1000 0] 0; 'ENDXL' [L L 0] [-1000 1000 0] [-1000 1000 0] 0; 'DISPDOF' [L L 0] [0 0 0] [0 0 0] 0; 'MID' [L/2 L/2 0] [max(node(:,3))-0.05 max(node(:,3))+0.05 0] [max(node(:,2))/2-0.05 max(node(:,2))/2+0.05 0] 0}; %DEFINE SPRINGS springs=[] %DEFINE CONTACT SURFACES, NODE SURFACES, KINEMATIC CONSTRAINTS,.... surface.type={} surface.type=[] surface.local=[]
424
surface.coord=[] surface.coupling={}; surface.interaction=[] surface.contact=[] surface.areadist=[] %ABAQUS INP FILE NAME jobname{count}=[SSMAnames{section_num} '_' num2str(i) '_' num2str(imptypes(j))]; %DEFINE ANALYSIS STEP step(1).stepinfo={'STEP 1,' 'nlgeom, INC=' [260]}; step(1).solutiontype='Static, Riks'; step(1).solutionsteps={'0.25, ,1e-10, 1'}; step(1).solutioncontrols={ }; step(1).boundarycon={'ENDXZERO' 2 3; 'ENDXL' 2 3; 'MID' 1 1} step(1).coupling=[] step(1).loads={'*Cload' end1cload(:,1) 1 end1cload(:,2)./2; '*Cload' end1cload(:,1)+200 1 end1cload(:,2)./2; '*Cload' end2cload(:,1) 1 end2cload(:,2)./2; '*Cload' end2cload(:,1)-200 1 end2cload(:,2)./2} step(1).outrequest={'*Output, field, frequency=10'; '*Element Output'; '1,3,5'; 'S,MISES'; '*Node Output'; 'U'; '*Node Print, NSET=DISPDOF, SUMMARY=NO'; 'U1,CF1'}; %WRITE ABAQUS INP FILE %this is the important function, you can use this in for loops to generate parameter studies jhabnl(L, node, elem, nele, end1load, end2load, hole, nodesetinfo, surface, nodeadd, step, jobname{count},matprops,imperfections,springs,sectionpoints) count=count+1 end end %CREATE BEAST BATCH FILE %Generates a linux batch file that will submit all of the parameter study %.inp files to the queue manager on the beast. ABQbeastscript(jobname,ones(length(jobname),1)*4,'cdmscript') %Run the script at the beast command line with: % bash cdmscript
425
Appendix B 10BABAQUS element-based elastic buckling results
This appendix contains the finite element plate model dimensions and ABAQUS critical
elastic buckling stress results (fcrl) used in the 1832HChapter 3 elastic buckling studies on
stiffened and unstiffened elements.
426
427
Stiffened element in uniaxial compression, transversely centered holes
hhole h Lhole S L t δholefcrl
in. in. in. in. in. in. in. ksi1 circular 1.50 15.00 1.50 20.00 100.00 0.0346 0.00 0.572 circular 1.50 7.50 1.50 20.00 100.00 0.0346 0.00 2.213 circular 1.50 5.00 1.50 20.00 100.00 0.0346 0.00 4.914 circular 1.50 3.75 1.50 20.00 100.00 0.0346 0.00 8.885 circular 1.50 3.00 1.50 20.00 100.00 0.0346 0.00 14.216 circular 1.50 2.50 1.50 20.00 100.00 0.0346 0.00 20.617 circular 1.50 2.14 1.50 20.00 100.00 0.0346 0.00 28.038 slotted 1.50 15.00 4.00 20.00 100.00 0.0346 0.00 0.539 slotted 1.50 7.50 4.00 20.00 100.00 0.0346 0.00 1.99
10 slotted 1.50 5.00 4.00 20.00 100.00 0.0346 0.00 4.3811 slotted 1.50 3.75 4.00 20.00 100.00 0.0346 0.00 8.0112 slotted 1.50 3.00 4.00 20.00 100.00 0.0346 0.00 13.3913 slotted 1.50 2.50 4.00 20.00 100.00 0.0346 0.00 20.6814 slotted 1.50 2.14 4.00 20.00 100.00 0.0346 0.00 28.0915 slotted 1.50 15.00 6.00 20.00 100.00 0.0346 0.00 0.4916 slotted 1.50 7.50 6.00 20.00 100.00 0.0346 0.00 1.8017 slotted 1.50 5.00 6.00 20.00 100.00 0.0346 0.00 3.9618 slotted 1.50 3.75 6.00 20.00 100.00 0.0346 0.00 7.2719 slotted 1.50 3.00 6.00 20.00 100.00 0.0346 0.00 12.4420 slotted 1.50 2.50 6.00 20.00 100.00 0.0346 0.00 20.7321 slotted 1.50 2.14 6.00 20.00 100.00 0.0346 0.00 28.1322 slotted 1.50 15.00 8.00 20.00 100.00 0.0346 0.00 0.4423 slotted 1.50 7.50 8.00 20.00 100.00 0.0346 0.00 1.6524 slotted 1.50 5.00 8.00 20.00 100.00 0.0346 0.00 3.6725 slotted 1.50 3.75 8.00 20.00 100.00 0.0346 0.00 6.8626 slotted 1.50 3.00 8.00 20.00 100.00 0.0346 0.00 12.0627 slotted 1.50 2.50 8.00 20.00 100.00 0.0346 0.00 20.7828 slotted 1.50 2.14 8.00 20.00 100.00 0.0346 0.00 28.2129 slotted 1.50 15.00 12.00 20.00 100.00 0.0346 0.00 0.3730 slotted 1.50 7.50 12.00 20.00 100.00 0.0346 0.00 1.4931 slotted 1.50 5.00 12.00 20.00 100.00 0.0346 0.00 3.3632 slotted 1.50 3.75 12.00 20.00 100.00 0.0346 0.00 6.5333 slotted 1.50 3.00 12.00 20.00 100.00 0.0346 0.00 11.8334 slotted 1.50 2.50 12.00 20.00 100.00 0.0346 0.00 20.8035 slotted 1.50 2.14 12.00 20.00 100.00 0.0346 0.00 22.8236 slotted 1.50 3.41 4.00 96.00 96.00 0.0346 0.00 9.9537 slotted 1.50 3.41 4.00 48.00 96.00 0.0346 0.00 9.9538 slotted 1.50 3.41 4.00 32.00 96.00 0.0346 0.00 9.9539 slotted 1.50 3.41 4.00 24.00 96.00 0.0346 0.00 9.9440 slotted 1.50 3.41 4.00 19.20 96.00 0.0346 0.00 9.9441 slotted 1.50 3.41 4.00 16.00 96.00 0.0346 0.00 9.9242 slotted 1.50 3.41 4.00 13.71 96.00 0.0346 0.00 9.9143 slotted 1.50 3.41 4.00 12.00 96.00 0.0346 0.00 9.8644 slotted 1.50 3.41 4.00 10.67 96.00 0.0346 0.00 9.9445 slotted 1.50 3.41 4.00 9.60 96.00 0.0346 0.00 9.7746 slotted 1.50 3.41 4.00 8.73 96.00 0.0346 0.00 9.7247 slotted 1.50 3.41 4.00 8.00 96.00 0.0346 0.00 9.8048 slotted 1.50 5.77 4.00 96.00 96.00 0.0346 0.00 3.2949 slotted 1.50 5.77 4.00 48.00 96.00 0.0346 0.00 3.2950 slotted 1.50 5.77 4.00 32.00 96.00 0.0346 0.00 3.2951 slotted 1.50 5.77 4.00 24.00 96.00 0.0346 0.00 3.2852 slotted 1.50 5.77 4.00 19.20 96.00 0.0346 0.00 3.2553 slotted 1.50 5.77 4.00 16.00 96.00 0.0346 0.00 3.2854 slotted 1.50 5.77 4.00 13.71 96.00 0.0346 0.00 3.1855 slotted 1.50 5.77 4.00 12.00 96.00 0.0346 0.00 3.1756 slotted 1.50 5.77 4.00 10.67 96.00 0.0346 0.00 3.2657 slotted 1.50 5.77 4.00 9.60 96.00 0.0346 0.00 3.2358 slotted 1.50 5.77 4.00 8.73 96.00 0.0346 0.00 3.0259 slotted 1.50 7.89 4.00 96.00 96.00 0.0346 0.00 1.8460 slotted 1.50 7.89 4.00 48.00 96.00 0.0346 0.00 1.8461 slotted 1.50 7.89 4.00 32.00 96.00 0.0346 0.00 1.8362 slotted 1.50 7.89 4.00 24.00 96.00 0.0346 0.00 1.8163 slotted 1.50 7.89 4.00 19.20 96.00 0.0346 0.00 1.8064 slotted 1.50 7.89 4.00 16.00 96.00 0.0346 0.00 1.7765 slotted 1.50 7.89 4.00 13.71 96.00 0.0346 0.00 1.8166 slotted 1.50 7.89 4.00 12.00 96.00 0.0346 0.00 1.8167 slotted 1.50 7.89 4.00 10.67 96.00 0.0346 0.00 1.6868 slotted 1.50 7.89 4.00 9.60 96.00 0.0346 0.00 1.7169 slotted 1.50 2.27 4.00 12.00 12.00 0.0346 0.00 26.4770 slotted 1.50 2.27 4.00 16.00 16.00 0.0346 0.00 25.4371 slotted 1.50 2.27 4.00 20.00 20.00 0.0346 0.00 25.0272 slotted 1.50 2.27 4.00 24.00 24.00 0.0346 0.00 24.8673 slotted 1.50 2.27 4.00 28.00 28.00 0.0346 0.00 24.8074 slotted 1.50 2.27 4.00 32.00 32.00 0.0346 0.00 24.7775 slotted 1.50 2.27 4.00 36.00 36.00 0.0346 0.00 24.75
Model number hole type
428
76 slotted 1.50 2.27 4.00 40.00 40.00 0.0346 0.00 24.7477 slotted 1.50 2.27 4.00 44.00 44.00 0.0346 0.00 24.7378 slotted 1.50 2.27 4.00 48.00 48.00 0.0346 0.00 24.7279 slotted 1.50 2.27 4.00 60.00 60.00 0.0346 0.00 24.6980 slotted 1.50 2.27 4.00 72.00 72.00 0.0346 0.00 24.6881 slotted 1.50 2.27 4.00 84.00 84.00 0.0346 0.00 24.6782 slotted 1.50 2.27 4.00 96.00 96.00 0.0346 0.00 24.6783 slotted 1.50 3.41 4.00 12.00 12.00 0.0346 0.00 9.8684 slotted 1.50 3.41 4.00 16.00 16.00 0.0346 0.00 9.9885 slotted 1.50 3.41 4.00 20.00 20.00 0.0346 0.00 9.9486 slotted 1.50 3.41 4.00 24.00 24.00 0.0346 0.00 9.9587 slotted 1.50 3.41 4.00 28.00 28.00 0.0346 0.00 9.9588 slotted 1.50 3.41 4.00 32.00 32.00 0.0346 0.00 9.9589 slotted 1.50 3.41 4.00 36.00 36.00 0.0346 0.00 9.9590 slotted 1.50 3.41 4.00 40.00 40.00 0.0346 0.00 9.9591 slotted 1.50 3.41 4.00 44.00 44.00 0.0346 0.00 9.9592 slotted 1.50 3.41 4.00 48.00 48.00 0.0346 0.00 9.9593 slotted 1.50 3.41 4.00 60.00 60.00 0.0346 0.00 9.9594 slotted 1.50 3.41 4.00 72.00 72.00 0.0346 0.00 9.9595 slotted 1.50 3.41 4.00 84.00 84.00 0.0346 0.00 9.9596 slotted 1.50 3.41 4.00 96.00 96.00 0.0346 0.00 9.9597 slotted 1.50 5.77 4.00 12.00 12.00 0.0346 0.00 3.5798 slotted 1.50 5.77 4.00 16.00 16.00 0.0346 0.00 3.3199 slotted 1.50 5.77 4.00 20.00 20.00 0.0346 0.00 3.26100 slotted 1.50 5.77 4.00 24.00 24.00 0.0346 0.00 3.30101 slotted 1.50 5.77 4.00 28.00 28.00 0.0346 0.00 3.30102 slotted 1.50 5.77 4.00 32.00 32.00 0.0346 0.00 3.29103 slotted 1.50 5.77 4.00 36.00 36.00 0.0346 0.00 3.29104 slotted 1.50 5.77 4.00 40.00 40.00 0.0346 0.00 3.29105 slotted 1.50 5.77 4.00 44.00 44.00 0.0346 0.00 3.29106 slotted 1.50 5.77 4.00 48.00 48.00 0.0346 0.00 3.29107 slotted 1.50 5.77 4.00 60.00 60.00 0.0346 0.00 3.29108 slotted 1.50 5.77 4.00 72.00 72.00 0.0346 0.00 3.29109 slotted 1.50 5.77 4.00 84.00 84.00 0.0346 0.00 3.29110 slotted 1.50 5.77 4.00 96.00 96.00 0.0346 0.00 3.29111 slotted 1.50 7.89 4.00 12.00 12.00 0.0346 0.00 1.81112 slotted 1.50 7.89 4.00 16.00 16.00 0.0346 0.00 2.02113 slotted 1.50 7.89 4.00 20.00 20.00 0.0346 0.00 1.87114 slotted 1.50 7.89 4.00 24.00 24.00 0.0346 0.00 1.81115 slotted 1.50 7.89 4.00 28.00 28.00 0.0346 0.00 1.83116 slotted 1.50 7.89 4.00 32.00 32.00 0.0346 0.00 1.85117 slotted 1.50 7.89 4.00 36.00 36.00 0.0346 0.00 1.84118 slotted 1.50 7.89 4.00 40.00 40.00 0.0346 0.00 1.84119 slotted 1.50 7.89 4.00 44.00 44.00 0.0346 0.00 1.84120 slotted 1.50 7.89 4.00 48.00 48.00 0.0346 0.00 1.84121 slotted 1.50 7.89 4.00 60.00 60.00 0.0346 0.00 1.84122 slotted 1.50 7.89 4.00 72.00 72.00 0.0346 0.00 1.84123 slotted 1.50 7.89 4.00 84.00 84.00 0.0346 0.00 1.84124 slotted 1.50 7.89 4.00 96.00 96.00 0.0346 0.00 1.84125 slotted 1.50 15.00 4.00 20.00 100.00 0.0692 0.00 2.14126 slotted 1.50 7.50 4.00 20.00 100.00 0.0692 0.00 7.96127 slotted 1.50 5.00 4.00 20.00 100.00 0.0692 0.00 17.45128 slotted 1.50 3.75 4.00 20.00 100.00 0.0692 0.00 31.73129 slotted 1.50 3.00 4.00 20.00 100.00 0.0692 0.00 52.49130 slotted 1.50 2.50 4.00 20.00 100.00 0.0692 0.00 82.03131 slotted 1.50 2.14 4.00 20.00 100.00 0.0692 0.00 111.50132 slotted 1.50 15.00 4.00 20.00 100.00 0.1038 0.00 4.81133 slotted 1.50 7.50 4.00 20.00 100.00 0.1038 0.00 17.87134 slotted 1.50 5.00 4.00 20.00 100.00 0.1038 0.00 39.05135 slotted 1.50 3.75 4.00 20.00 100.00 0.1038 0.00 70.42136 slotted 1.50 3.00 4.00 20.00 100.00 0.1038 0.00 114.77137 slotted 1.50 2.50 4.00 20.00 100.00 0.1038 0.00 178.44138 slotted 1.50 2.14 4.00 20.00 100.00 0.1038 0.00 242.93139 square 1.50 15.00 1.50 20.00 100.00 0.0346 0.00 0.56140 square 1.50 7.50 1.50 20.00 100.00 0.0346 0.00 2.21141 square 1.50 5.00 1.50 20.00 100.00 0.0346 0.00 4.95142 square 1.50 3.75 1.50 20.00 100.00 0.0346 0.00 9.02143 square 1.50 3.00 1.50 20.00 100.00 0.0346 0.00 14.32144 square 1.50 2.50 1.50 20.00 100.00 0.0346 0.00 20.62145 square 1.50 2.14 1.50 20.00 100.00 0.0346 0.00 27.99
429
Stiffened element in uniaxial compression, offset holes
hhole h Lhole S L t δhole fcrl
in. in. in. in. in. in. in. ksi1 slotted 1.50 15.00 4.00 20.00 100.00 0.0346 0.00 0.532 slotted 1.50 15.00 4.00 20.00 100.00 0.0346 0.94 0.543 slotted 1.50 15.00 4.00 20.00 100.00 0.0346 1.88 0.544 slotted 1.50 15.00 4.00 20.00 100.00 0.0346 2.81 0.545 slotted 1.50 15.00 4.00 20.00 100.00 0.0346 3.75 0.546 slotted 1.50 15.00 4.00 20.00 100.00 0.0346 4.69 0.547 slotted 1.50 15.00 4.00 20.00 100.00 0.0346 5.63 0.548 slotted 1.50 7.50 4.00 20.00 100.00 0.0346 0.00 1.999 slotted 1.50 7.50 4.00 20.00 100.00 0.0346 0.47 1.9810 slotted 1.50 7.50 4.00 20.00 100.00 0.0346 0.94 1.9611 slotted 1.50 7.50 4.00 20.00 100.00 0.0346 1.41 1.9612 slotted 1.50 7.50 4.00 20.00 100.00 0.0346 1.88 1.9113 slotted 1.50 7.50 4.00 20.00 100.00 0.0346 2.34 1.8514 slotted 1.50 5.00 4.00 20.00 100.00 0.0346 0.00 4.3815 slotted 1.50 5.00 4.00 20.00 100.00 0.0346 0.31 4.1716 slotted 1.50 5.00 4.00 20.00 100.00 0.0346 0.63 3.7917 slotted 1.50 5.00 4.00 20.00 100.00 0.0346 0.94 3.7118 slotted 1.50 5.00 4.00 20.00 100.00 0.0346 1.25 3.3819 slotted 1.50 3.75 4.00 20.00 100.00 0.0346 0.07 7.8120 slotted 1.50 3.75 4.00 20.00 100.00 0.0346 0.14 7.4321 slotted 1.50 3.75 4.00 20.00 100.00 0.0346 0.20 7.0322 slotted 1.50 3.75 4.00 20.00 100.00 0.0346 0.27 6.6523 slotted 1.50 3.75 4.00 20.00 100.00 0.0346 0.34 6.3124 slotted 1.50 3.75 4.00 20.00 100.00 0.0346 0.41 5.9925 slotted 1.50 3.75 4.00 20.00 100.00 0.0346 0.47 5.7026 slotted 1.50 3.00 4.00 20.00 100.00 0.0346 0.05 12.6527 slotted 1.50 3.00 4.00 20.00 100.00 0.0346 0.09 11.7628 slotted 1.50 3.00 4.00 20.00 100.00 0.0346 0.14 10.9529 slotted 1.50 3.00 4.00 20.00 100.00 0.0346 0.18 10.2330 slotted 1.50 3.00 4.00 20.00 100.00 0.0346 0.23 9.5931 slotted 1.50 3.00 4.00 20.00 100.00 0.0346 0.27 9.0132 slotted 1.50 3.00 4.00 20.00 100.00 0.0346 0.32 8.4933 slotted 1.50 2.50 4.00 20.00 100.00 0.0346 0.03 20.1234 slotted 1.50 2.50 4.00 20.00 100.00 0.0346 0.06 18.5535 slotted 1.50 2.50 4.00 20.00 100.00 0.0346 0.09 17.1536 slotted 1.50 2.50 4.00 20.00 100.00 0.0346 0.12 15.9137 slotted 1.50 2.50 4.00 20.00 100.00 0.0346 0.15 14.8038 slotted 1.50 2.50 4.00 20.00 100.00 0.0346 0.18 13.8139 slotted 1.50 2.50 4.00 20.00 100.00 0.0346 0.21 12.9140 slotted 1.50 2.14 4.00 20.00 100.00 0.0346 0.02 28.5441 slotted 1.50 2.14 4.00 20.00 100.00 0.0346 0.04 28.5442 slotted 1.50 2.14 4.00 20.00 100.00 0.0346 0.06 27.9943 slotted 1.50 2.14 4.00 20.00 100.00 0.0346 0.08 25.8844 slotted 1.50 2.14 4.00 20.00 100.00 0.0346 0.10 23.9845 slotted 1.50 2.14 4.00 20.00 100.00 0.0346 0.12 22.2646 slotted 1.50 2.14 4.00 20.00 100.00 0.0346 0.14 20.72
Model number hole type
430
Stiffened element in bending (Y=0.50h), transversely centered holes
hhole h Lhole S L t δhole Y fcrl
in. in. in. in. in. in. in. in. ksi1 slotted 1.50 15.00 4.00 20.00 100.00 0.0346 0.00 7.50 3.262 slotted 1.50 15.00 6.00 20.00 100.00 0.0346 0.00 7.50 3.063 slotted 1.50 15.00 8.00 20.00 100.00 0.0346 0.00 7.50 2.784 slotted 1.50 15.00 12.00 20.00 100.00 0.0346 0.00 7.50 2.235 slotted 1.50 7.50 4.00 20.00 100.00 0.0346 0.00 3.75 10.636 slotted 1.50 7.50 6.00 20.00 100.00 0.0346 0.00 3.75 8.277 slotted 1.50 7.50 8.00 20.00 100.00 0.0346 0.00 3.75 6.778 slotted 1.50 7.50 12.00 20.00 100.00 0.0346 0.00 3.75 5.289 slotted 1.50 5.00 4.00 20.00 100.00 0.0346 0.00 2.50 17.97
10 slotted 1.50 5.00 6.00 20.00 100.00 0.0346 0.00 2.50 13.8911 slotted 1.50 5.00 8.00 20.00 100.00 0.0346 0.00 2.50 11.9712 slotted 1.50 5.00 12.00 20.00 100.00 0.0346 0.00 2.50 10.4013 slotted 1.50 3.75 4.00 20.00 100.00 0.0346 0.00 1.88 27.1614 slotted 1.50 3.75 6.00 20.00 100.00 0.0346 0.00 1.88 22.4215 slotted 1.50 3.75 8.00 20.00 100.00 0.0346 0.00 1.88 20.5116 slotted 1.50 3.75 12.00 20.00 100.00 0.0346 0.00 1.88 19.1117 slotted 1.50 3.00 4.00 20.00 100.00 0.0346 0.00 1.50 41.1418 slotted 1.50 3.00 6.00 20.00 100.00 0.0346 0.00 1.50 36.4719 slotted 1.50 3.00 8.00 20.00 100.00 0.0346 0.00 1.50 34.8520 slotted 1.50 3.00 12.00 20.00 100.00 0.0346 0.00 1.50 33.8121 slotted 1.50 2.50 4.00 20.00 100.00 0.0346 0.00 1.25 64.4222 slotted 1.50 2.50 6.00 20.00 100.00 0.0346 0.00 1.25 60.5923 slotted 1.50 2.50 8.00 20.00 100.00 0.0346 0.00 1.25 59.5224 slotted 1.50 2.50 12.00 20.00 100.00 0.0346 0.00 1.25 59.0725 slotted 1.50 2.14 4.00 20.00 100.00 0.0346 0.00 1.07 106.3026 slotted 1.50 2.14 6.00 20.00 100.00 0.0346 0.00 1.07 104.3927 slotted 1.50 2.14 8.00 20.00 100.00 0.0346 0.00 1.07 104.2528 slotted 1.50 2.14 12.00 20.00 100.00 0.0346 0.00 1.07 58.03
Model number hole type
431
Stiffened element in bending (Y=0.50h), offset holes hhole h Lhole S L t δhole Y fcrl
in. in. in. in. in. in. in. in. ksi1 slotted 1.50 15.00 4.00 20.00 100.00 0.0346 0.00 7.50 3.262 slotted 1.50 15.00 4.00 20.00 100.00 0.0346 -0.94 7.50 3.343 slotted 1.50 15.00 4.00 20.00 100.00 0.0346 -1.88 7.50 3.384 slotted 1.50 15.00 4.00 20.00 100.00 0.0346 -2.81 7.50 3.415 slotted 1.50 15.00 4.00 20.00 100.00 0.0346 -3.75 7.50 3.446 slotted 1.50 15.00 4.00 20.00 100.00 0.0346 -4.69 7.50 3.457 slotted 1.50 15.00 4.00 20.00 100.00 0.0346 -5.63 7.50 3.448 slotted 1.50 7.50 4.00 20.00 100.00 0.0346 0.00 3.75 10.639 slotted 1.50 7.50 4.00 20.00 100.00 0.0346 -0.47 3.75 12.1010 slotted 1.50 7.50 4.00 20.00 100.00 0.0346 -0.94 3.75 13.3511 slotted 1.50 7.50 4.00 20.00 100.00 0.0346 -1.41 3.75 13.8012 slotted 1.50 7.50 4.00 20.00 100.00 0.0346 -1.88 3.75 13.9713 slotted 1.50 7.50 4.00 20.00 100.00 0.0346 -2.34 3.75 13.9214 slotted 1.50 5.00 4.00 20.00 100.00 0.0346 0.00 2.50 17.9615 slotted 1.50 5.00 4.00 20.00 100.00 0.0346 -0.31 2.50 21.2516 slotted 1.50 5.00 4.00 20.00 100.00 0.0346 -0.63 2.50 27.8617 slotted 1.50 5.00 4.00 20.00 100.00 0.0346 -0.94 2.50 31.0318 slotted 1.50 5.00 4.00 20.00 100.00 0.0346 -1.25 2.50 31.1819 slotted 1.50 3.75 4.00 20.00 100.00 0.0346 -0.07 1.88 27.7420 slotted 1.50 3.75 4.00 20.00 100.00 0.0346 -0.14 1.88 28.6721 slotted 1.50 3.75 4.00 20.00 100.00 0.0346 -0.20 1.88 30.0522 slotted 1.50 3.75 4.00 20.00 100.00 0.0346 -0.27 1.88 32.0123 slotted 1.50 3.75 4.00 20.00 100.00 0.0346 -0.34 1.88 34.7424 slotted 1.50 3.75 4.00 20.00 100.00 0.0346 -0.41 1.88 38.5725 slotted 1.50 3.75 4.00 20.00 100.00 0.0346 -0.47 1.88 43.9126 slotted 1.50 3.00 4.00 20.00 100.00 0.0346 -0.05 1.50 41.0927 slotted 1.50 3.00 4.00 20.00 100.00 0.0346 -0.09 1.50 41.4528 slotted 1.50 3.00 4.00 20.00 100.00 0.0346 -0.14 1.50 42.2829 slotted 1.50 3.00 4.00 20.00 100.00 0.0346 -0.18 1.50 43.6630 slotted 1.50 3.00 4.00 20.00 100.00 0.0346 -0.23 1.50 45.7631 slotted 1.50 3.00 4.00 20.00 100.00 0.0346 -0.27 1.50 48.8732 slotted 1.50 3.00 4.00 20.00 100.00 0.0346 -0.32 1.50 53.4933 slotted 1.50 2.50 4.00 20.00 100.00 0.0346 -0.03 1.25 63.1734 slotted 1.50 2.50 4.00 20.00 100.00 0.0346 -0.06 1.25 62.4835 slotted 1.50 2.50 4.00 20.00 100.00 0.0346 -0.09 1.25 62.3036 slotted 1.50 2.50 4.00 20.00 100.00 0.0346 -0.12 1.25 62.6837 slotted 1.50 2.50 4.00 20.00 100.00 0.0346 -0.15 1.25 63.7338 slotted 1.50 2.50 4.00 20.00 100.00 0.0346 -0.18 1.25 65.6239 slotted 1.50 2.50 4.00 20.00 100.00 0.0346 -0.21 1.25 68.6940 slotted 1.50 2.14 4.00 20.00 100.00 0.0346 -0.02 1.07 102.3341 slotted 1.50 2.14 4.00 20.00 100.00 0.0346 -0.04 1.07 99.9442 slotted 1.50 2.14 4.00 20.00 100.00 0.0346 -0.06 1.07 98.1243 slotted 1.50 2.14 4.00 20.00 100.00 0.0346 -0.08 1.07 96.9244 slotted 1.50 2.14 4.00 20.00 100.00 0.0346 -0.10 1.07 96.4045 slotted 1.50 2.14 4.00 20.00 100.00 0.0346 -0.12 1.07 96.6946 slotted 1.50 2.14 4.00 20.00 100.00 0.0346 -0.14 1.07 97.9947 slotted 1.50 15.00 4.00 20.00 100.00 0.0346 0.00 7.50 3.2648 slotted 1.50 15.00 4.00 20.00 100.00 0.0346 0.94 7.50 3.1449 slotted 1.50 15.00 4.00 20.00 100.00 0.0346 1.88 7.50 3.0450 slotted 1.50 15.00 4.00 20.00 100.00 0.0346 2.81 7.50 2.9651 slotted 1.50 15.00 4.00 20.00 100.00 0.0346 3.75 7.50 2.9152 slotted 1.50 15.00 4.00 20.00 100.00 0.0346 4.69 7.50 2.8153 slotted 1.50 15.00 4.00 20.00 100.00 0.0346 5.63 7.50 2.5754 slotted 1.50 7.50 4.00 20.00 100.00 0.0346 0.00 3.75 10.6355 slotted 1.50 7.50 4.00 20.00 100.00 0.0346 0.47 3.75 9.7256 slotted 1.50 7.50 4.00 20.00 100.00 0.0346 0.94 3.75 9.5357 slotted 1.50 7.50 4.00 20.00 100.00 0.0346 1.41 3.75 10.1758 slotted 1.50 7.50 4.00 20.00 100.00 0.0346 1.88 3.75 10.8759 slotted 1.50 7.50 4.00 20.00 100.00 0.0346 2.34 3.75 7.3660 slotted 1.50 5.00 4.00 20.00 100.00 0.0346 0.00 2.50 17.9661 slotted 1.50 5.00 4.00 20.00 100.00 0.0346 0.31 2.50 16.9562 slotted 1.50 5.00 4.00 20.00 100.00 0.0346 0.63 2.50 17.8863 slotted 1.50 5.00 4.00 20.00 100.00 0.0346 0.94 2.50 21.7864 slotted 1.50 5.00 4.00 20.00 100.00 0.0346 1.25 2.50 19.2965 slotted 1.50 3.75 4.00 20.00 100.00 0.0346 0.07 1.88 26.9066 slotted 1.50 3.75 4.00 20.00 100.00 0.0346 0.14 1.88 26.9367 slotted 1.50 3.75 4.00 20.00 100.00 0.0346 0.20 1.88 27.2568 slotted 1.50 3.75 4.00 20.00 100.00 0.0346 0.27 1.88 27.8869 slotted 1.50 3.75 4.00 20.00 100.00 0.0346 0.34 1.88 28.8770 slotted 1.50 3.75 4.00 20.00 100.00 0.0346 0.41 1.88 30.2871 slotted 1.50 3.75 4.00 20.00 100.00 0.0346 0.47 1.88 32.2272 slotted 1.50 3.00 4.00 20.00 100.00 0.0346 0.05 1.50 41.5473 slotted 1.50 3.00 4.00 20.00 100.00 0.0346 0.09 1.50 42.3474 slotted 1.50 3.00 4.00 20.00 100.00 0.0346 0.14 1.50 43.5575 slotted 1.50 3.00 4.00 20.00 100.00 0.0346 0.18 1.50 45.2376 slotted 1.50 3.00 4.00 20.00 100.00 0.0346 0.23 1.50 47.4477 slotted 1.50 3.00 4.00 20.00 100.00 0.0346 0.27 1.50 50.3178 slotted 1.50 3.00 4.00 20.00 100.00 0.0346 0.32 1.50 54.0279 slotted 1.50 2.50 4.00 20.00 100.00 0.0346 0.03 1.25 66.0280 slotted 1.50 2.50 4.00 20.00 100.00 0.0346 0.06 1.25 68.2181 slotted 1.50 2.50 4.00 20.00 100.00 0.0346 0.09 1.25 70.9982 slotted 1.50 2.50 4.00 20.00 100.00 0.0346 0.12 1.25 74.4283 slotted 1.50 2.50 4.00 20.00 100.00 0.0346 0.15 1.25 78.6384 slotted 1.50 2.50 4.00 20.00 100.00 0.0346 0.18 1.25 83.6985 slotted 1.50 2.50 4.00 20.00 100.00 0.0346 0.21 1.25 89.6486 slotted 1.50 2.14 4.00 20.00 100.00 0.0346 0.02 1.07 108.6687 slotted 1.50 2.14 4.00 20.00 100.00 0.0346 0.04 1.07 112.4688 slotted 1.50 2.14 4.00 20.00 100.00 0.0346 0.06 1.07 116.4489 slotted 1.50 2.14 4.00 20.00 100.00 0.0346 0.08 1.07 116.3590 slotted 1.50 2.14 4.00 20.00 100.00 0.0346 0.10 1.07 112.5491 slotted 1.50 2.14 4.00 20.00 100.00 0.0346 0.12 1.07 107.2492 slotted 1.50 2.14 4.00 20.00 100.00 0.0346 0.14 1.07 100.88
Model number hole type
432
Stiffened element in bending (Y=0.75h), offset holes hhole h Lhole S L t δhole Y fcrl
in. in. in. in. in. in. in. in. ksi1 slotted 1.50 15.00 4.00 20.00 100.00 0.0346 0.00 11.25 1.482 slotted 1.50 15.00 4.00 20.00 100.00 0.0346 -0.94 11.25 1.493 slotted 1.50 15.00 4.00 20.00 100.00 0.0346 -1.88 11.25 1.504 slotted 1.50 15.00 4.00 20.00 100.00 0.0346 -2.81 11.25 1.515 slotted 1.50 15.00 4.00 20.00 100.00 0.0346 -3.75 11.25 1.536 slotted 1.50 15.00 4.00 20.00 100.00 0.0346 -4.69 11.25 1.547 slotted 1.50 15.00 4.00 20.00 100.00 0.0346 -5.63 11.25 1.568 slotted 1.50 7.50 4.00 20.00 100.00 0.0346 0.00 5.63 5.209 slotted 1.50 7.50 4.00 20.00 100.00 0.0346 -0.47 5.63 5.2210 slotted 1.50 7.50 4.00 20.00 100.00 0.0346 -0.94 5.63 5.3311 slotted 1.50 7.50 4.00 20.00 100.00 0.0346 -1.41 5.63 5.5212 slotted 1.50 7.50 4.00 20.00 100.00 0.0346 -1.88 5.63 5.7713 slotted 1.50 7.50 4.00 20.00 100.00 0.0346 -2.34 5.63 6.0414 slotted 1.50 5.00 4.00 20.00 100.00 0.0346 0.00 3.75 9.9715 slotted 1.50 5.00 4.00 20.00 100.00 0.0346 -0.31 3.75 9.6216 slotted 1.50 5.00 4.00 20.00 100.00 0.0346 -0.63 3.75 9.6617 slotted 1.50 5.00 4.00 20.00 100.00 0.0346 -0.94 3.75 10.0918 slotted 1.50 5.00 4.00 20.00 100.00 0.0346 -1.25 3.75 11.0719 slotted 1.50 3.75 4.00 20.00 100.00 0.0346 -0.07 2.81 15.5320 slotted 1.50 3.75 4.00 20.00 100.00 0.0346 -0.14 2.81 15.1421 slotted 1.50 3.75 4.00 20.00 100.00 0.0346 -0.20 2.81 14.8322 slotted 1.50 3.75 4.00 20.00 100.00 0.0346 -0.27 2.81 14.5923 slotted 1.50 3.75 4.00 20.00 100.00 0.0346 -0.34 2.81 14.4224 slotted 1.50 3.75 4.00 20.00 100.00 0.0346 -0.41 2.81 14.3025 slotted 1.50 3.75 4.00 20.00 100.00 0.0346 -0.47 2.81 14.2426 slotted 1.50 3.00 4.00 20.00 100.00 0.0346 -0.05 2.25 23.8627 slotted 1.50 3.00 4.00 20.00 100.00 0.0346 -0.09 2.25 23.0128 slotted 1.50 3.00 4.00 20.00 100.00 0.0346 -0.14 2.25 22.2929 slotted 1.50 3.00 4.00 20.00 100.00 0.0346 -0.18 2.25 21.6630 slotted 1.50 3.00 4.00 20.00 100.00 0.0346 -0.23 2.25 21.1331 slotted 1.50 3.00 4.00 20.00 100.00 0.0346 -0.27 2.25 20.6832 slotted 1.50 3.00 4.00 20.00 100.00 0.0346 -0.32 2.25 20.3233 slotted 1.50 2.50 4.00 20.00 100.00 0.0346 -0.03 1.88 37.4434 slotted 1.50 2.50 4.00 20.00 100.00 0.0346 -0.06 1.88 35.8335 slotted 1.50 2.50 4.00 20.00 100.00 0.0346 -0.09 1.88 34.4036 slotted 1.50 2.50 4.00 20.00 100.00 0.0346 -0.12 1.88 33.1437 slotted 1.50 2.50 4.00 20.00 100.00 0.0346 -0.15 1.88 32.0138 slotted 1.50 2.50 4.00 20.00 100.00 0.0346 -0.18 1.88 31.0139 slotted 1.50 2.50 4.00 20.00 100.00 0.0346 -0.21 1.88 30.1140 slotted 1.50 2.14 4.00 20.00 100.00 0.0346 -0.02 1.61 61.2741 slotted 1.50 2.14 4.00 20.00 100.00 0.0346 -0.04 1.61 58.4742 slotted 1.50 2.14 4.00 20.00 100.00 0.0346 -0.06 1.61 55.9243 slotted 1.50 2.14 4.00 20.00 100.00 0.0346 -0.08 1.61 53.6244 slotted 1.50 2.14 4.00 20.00 100.00 0.0346 -0.10 1.61 51.5545 slotted 1.50 2.14 4.00 20.00 100.00 0.0346 -0.12 1.61 49.5846 slotted 1.50 2.14 4.00 20.00 100.00 0.0346 -0.14 1.61 47.8147 slotted 1.50 15.00 4.00 20.00 100.00 0.0346 0.00 11.25 1.4848 slotted 1.50 15.00 4.00 20.00 100.00 0.0346 0.94 11.25 1.4749 slotted 1.50 15.00 4.00 20.00 100.00 0.0346 1.88 11.25 1.4550 slotted 1.50 15.00 4.00 20.00 100.00 0.0346 2.81 11.25 1.4451 slotted 1.50 15.00 4.00 20.00 100.00 0.0346 3.75 11.25 1.4352 slotted 1.50 15.00 4.00 20.00 100.00 0.0346 4.69 11.25 1.4053 slotted 1.50 15.00 4.00 20.00 100.00 0.0346 5.63 11.25 1.3654 slotted 1.50 7.50 4.00 20.00 100.00 0.0346 0.00 5.63 5.1955 slotted 1.50 7.50 4.00 20.00 100.00 0.0346 0.47 5.63 5.2556 slotted 1.50 7.50 4.00 20.00 100.00 0.0346 0.94 5.63 5.3157 slotted 1.50 7.50 4.00 20.00 100.00 0.0346 1.41 5.63 5.1758 slotted 1.50 7.50 4.00 20.00 100.00 0.0346 1.88 5.63 4.6259 slotted 1.50 7.50 4.00 20.00 100.00 0.0346 2.34 5.63 3.8660 slotted 1.50 5.00 4.00 20.00 100.00 0.0346 0.00 3.75 9.9761 slotted 1.50 5.00 4.00 20.00 100.00 0.0346 0.31 3.75 10.7962 slotted 1.50 5.00 4.00 20.00 100.00 0.0346 0.63 3.75 11.8363 slotted 1.50 5.00 4.00 20.00 100.00 0.0346 0.94 3.75 10.0264 slotted 1.50 5.00 4.00 20.00 100.00 0.0346 1.25 3.75 7.2465 slotted 1.50 3.75 4.00 20.00 100.00 0.0346 0.07 2.81 16.5866 slotted 1.50 3.75 4.00 20.00 100.00 0.0346 0.14 2.81 17.2767 slotted 1.50 3.75 4.00 20.00 100.00 0.0346 0.20 2.81 18.1168 slotted 1.50 3.75 4.00 20.00 100.00 0.0346 0.27 2.81 19.1369 slotted 1.50 3.75 4.00 20.00 100.00 0.0346 0.34 2.81 20.3570 slotted 1.50 3.75 4.00 20.00 100.00 0.0346 0.41 2.81 21.7571 slotted 1.50 3.75 4.00 20.00 100.00 0.0346 0.47 2.81 21.7972 slotted 1.50 3.00 4.00 20.00 100.00 0.0346 0.05 2.25 25.9973 slotted 1.50 3.00 4.00 20.00 100.00 0.0346 0.09 2.25 27.3474 slotted 1.50 3.00 4.00 20.00 100.00 0.0346 0.14 2.25 28.9275 slotted 1.50 3.00 4.00 20.00 100.00 0.0346 0.18 2.25 30.8276 slotted 1.50 3.00 4.00 20.00 100.00 0.0346 0.23 2.25 33.0977 slotted 1.50 3.00 4.00 20.00 100.00 0.0346 0.27 2.25 35.8478 slotted 1.50 3.00 4.00 20.00 100.00 0.0346 0.32 2.25 38.2179 slotted 1.50 2.50 4.00 20.00 100.00 0.0346 0.03 1.88 41.3580 slotted 1.50 2.50 4.00 20.00 100.00 0.0346 0.06 1.88 43.7581 slotted 1.50 2.50 4.00 20.00 100.00 0.0346 0.09 1.88 46.5282 slotted 1.50 2.50 4.00 20.00 100.00 0.0346 0.12 1.88 49.7483 slotted 1.50 2.50 4.00 20.00 100.00 0.0346 0.15 1.88 53.5184 slotted 1.50 2.50 4.00 20.00 100.00 0.0346 0.18 1.88 56.6885 slotted 1.50 2.50 4.00 20.00 100.00 0.0346 0.21 1.88 56.6686 slotted 1.50 2.14 4.00 20.00 100.00 0.0346 0.02 1.61 67.6787 slotted 1.50 2.14 4.00 20.00 100.00 0.0346 0.04 1.61 71.2588 slotted 1.50 2.14 4.00 20.00 100.00 0.0346 0.06 1.61 74.9689 slotted 1.50 2.14 4.00 20.00 100.00 0.0346 0.08 1.61 76.6390 slotted 1.50 2.14 4.00 20.00 100.00 0.0346 0.10 1.61 75.1191 slotted 1.50 2.14 4.00 20.00 100.00 0.0346 0.12 1.61 72.3892 slotted 1.50 2.14 4.00 20.00 100.00 0.0346 0.14 1.61 68.78
Model number hole type
433
Unstiffened element in uniaxial compression, transversely centered holes
hhole h Lhole S L t δhole fcrl
in. in. in. in. in. in. in. ksi1 ang. slotted 1.50 15.00 4.00 20.00 100.00 0.0346 0.00 0.062 ang. slotted 1.50 7.50 4.00 20.00 100.00 0.0346 0.00 0.233 ang. slotted 1.50 5.00 4.00 20.00 100.00 0.0346 0.00 0.484 ang. slotted 1.50 3.75 4.00 20.00 100.00 0.0346 0.00 0.765 ang. slotted 1.50 3.00 4.00 20.00 100.00 0.0346 0.00 0.986 ang. slotted 1.50 2.50 4.00 20.00 100.00 0.0346 0.00 1.087 ang. slotted 1.50 2.14 4.00 20.00 100.00 0.0346 0.00 1.038 slotted 1.50 15.00 4.00 20.00 100.00 0.0346 0.00 0.069 slotted 1.50 7.50 4.00 20.00 100.00 0.0346 0.00 0.2310 slotted 1.50 5.00 4.00 20.00 100.00 0.0346 0.00 0.4811 slotted 1.50 3.75 4.00 20.00 100.00 0.0346 0.00 0.7612 slotted 1.50 3.00 4.00 20.00 100.00 0.0346 0.00 0.9813 slotted 1.50 2.50 4.00 20.00 100.00 0.0346 0.00 1.0814 slotted 1.50 2.14 4.00 20.00 100.00 0.0346 0.00 1.0315 slotted 1.50 15.00 6.00 20.00 100.00 0.0346 0.00 0.0616 slotted 1.50 7.50 6.00 20.00 100.00 0.0346 0.00 0.2217 slotted 1.50 5.00 6.00 20.00 100.00 0.0346 0.00 0.4518 slotted 1.50 3.75 6.00 20.00 100.00 0.0346 0.00 0.6719 slotted 1.50 3.00 6.00 20.00 100.00 0.0346 0.00 0.8120 slotted 1.50 2.50 6.00 20.00 100.00 0.0346 0.00 0.8521 slotted 1.50 2.14 6.00 20.00 100.00 0.0346 0.00 0.7722 slotted 1.50 15.00 8.00 20.00 100.00 0.0346 0.00 0.0623 slotted 1.50 7.50 8.00 20.00 100.00 0.0346 0.00 0.2124 slotted 1.50 5.00 8.00 20.00 100.00 0.0346 0.00 0.4325 slotted 1.50 3.75 8.00 20.00 100.00 0.0346 0.00 0.6126 slotted 1.50 3.00 8.00 20.00 100.00 0.0346 0.00 0.6327 slotted 1.50 2.50 8.00 20.00 100.00 0.0346 0.00 0.5828 slotted 1.50 2.14 8.00 20.00 100.00 0.0346 0.00 0.4929 slotted 1.50 15.00 12.00 20.00 100.00 0.0346 0.00 0.0630 slotted 1.50 7.50 12.00 20.00 100.00 0.0346 0.00 0.2031 slotted 1.50 5.00 12.00 20.00 100.00 0.0346 0.00 0.3132 slotted 1.50 3.75 12.00 20.00 100.00 0.0346 0.00 0.3433 slotted 1.50 3.00 12.00 20.00 100.00 0.0346 0.00 0.3234 slotted 1.50 2.50 12.00 20.00 100.00 0.0346 0.00 0.2835 slotted 1.50 2.14 12.00 20.00 100.00 0.0346 0.00 0.2336 slotted 1.50 15.00 4.00 20.00 100.00 0.0692 0.00 0.2537 slotted 1.50 7.50 4.00 20.00 100.00 0.0692 0.00 0.9138 slotted 1.50 5.00 4.00 20.00 100.00 0.0692 0.00 1.9339 slotted 1.50 3.75 4.00 20.00 100.00 0.0692 0.00 3.0440 slotted 1.50 3.00 4.00 20.00 100.00 0.0692 0.00 3.8841 slotted 1.50 2.50 4.00 20.00 100.00 0.0692 0.00 4.2242 slotted 1.50 2.14 4.00 20.00 100.00 0.0692 0.00 3.9643 slotted 1.50 15.00 4.00 20.00 100.00 0.1038 0.00 0.5544 slotted 1.50 7.50 4.00 20.00 100.00 0.1038 0.00 2.0545 slotted 1.50 5.00 4.00 20.00 100.00 0.1038 0.00 4.3346 slotted 1.50 3.75 4.00 20.00 100.00 0.1038 0.00 6.7947 slotted 1.50 3.00 4.00 20.00 100.00 0.1038 0.00 8.5948 slotted 1.50 2.50 4.00 20.00 100.00 0.1038 0.00 9.2349 slotted 1.50 2.14 4.00 20.00 100.00 0.1038 0.00 8.5250 slotted 1.50 7.89 4.00 96.00 96.00 0.0346 0.00 0.2251 slotted 1.50 7.89 4.00 48.00 96.00 0.0346 0.00 0.2252 slotted 1.50 7.89 4.00 32.00 96.00 0.0346 0.00 0.2253 slotted 1.50 7.89 4.00 24.00 96.00 0.0346 0.00 0.2154 slotted 1.50 7.89 4.00 16.00 96.00 0.0346 0.00 0.2155 slotted 1.50 7.89 4.00 12.00 96.00 0.0346 0.00 0.2056 slotted 1.50 7.89 4.00 8.00 96.00 0.0346 0.00 0.1957 slotted 1.50 5.77 4.00 96.00 96.00 0.0346 0.00 0.4058 slotted 1.50 5.77 4.00 48.00 96.00 0.0346 0.00 0.3859 slotted 1.50 5.77 4.00 32.00 96.00 0.0346 0.00 0.3860 slotted 1.50 5.77 4.00 24.00 96.00 0.0346 0.00 0.3861 slotted 1.50 5.77 4.00 16.00 96.00 0.0346 0.00 0.3762 slotted 1.50 5.77 4.00 12.00 96.00 0.0346 0.00 0.3663 slotted 1.50 5.77 4.00 8.00 96.00 0.0346 0.00 0.3464 slotted 1.50 3.41 4.00 96.00 96.00 0.0346 0.00 0.9165 slotted 1.50 3.41 4.00 48.00 96.00 0.0346 0.00 0.8766 slotted 1.50 3.41 4.00 32.00 96.00 0.0346 0.00 0.8767 slotted 1.50 3.41 4.00 24.00 96.00 0.0346 0.00 0.8768 slotted 1.50 3.41 4.00 16.00 96.00 0.0346 0.00 0.8669 slotted 1.50 3.41 4.00 12.00 96.00 0.0346 0.00 0.8470 slotted 1.50 3.41 4.00 8.00 96.00 0.0346 0.00 0.7871 slotted 1.50 2.27 4.00 96.00 96.00 0.0346 0.00 1.1672 slotted 1.50 2.27 4.00 48.00 96.00 0.0346 0.00 1.0873 slotted 1.50 2.27 4.00 32.00 96.00 0.0346 0.00 1.0874 slotted 1.50 2.27 4.00 24.00 96.00 0.0346 0.00 1.0875 slotted 1.50 2.27 4.00 16.00 96.00 0.0346 0.00 1.0876 slotted 1.50 2.27 4.00 12.00 96.00 0.0346 0.00 1.0877 slotted 1.50 2.27 4.00 8.00 96.00 0.0346 0.00 1.0678 square 1.50 15.00 1.50 20.00 100.00 0.0346 0.00 0.0679 square 1.50 7.50 1.50 20.00 100.00 0.0346 0.00 0.2480 square 1.50 5.00 1.50 20.00 100.00 0.0346 0.00 0.5281 square 1.50 3.75 1.50 20.00 100.00 0.0346 0.00 0.8882 square 1.50 3.00 1.50 20.00 100.00 0.0346 0.00 1.2683 square 1.50 2.50 1.50 20.00 100.00 0.0346 0.00 1.5784 square 1.50 2.14 1.50 20.00 100.00 0.0346 0.00 1.7185 circular 1.50 15.00 1.50 20.00 100.00 0.0346 0.00 0.0686 circular 1.50 7.50 1.50 20.00 100.00 0.0346 0.00 0.2487 circular 1.50 5.00 1.50 20.00 100.00 0.0346 0.00 0.5288 circular 1.50 3.75 1.50 20.00 100.00 0.0346 0.00 0.9189 circular 1.50 3.00 1.50 20.00 100.00 0.0346 0.00 1.3490 circular 1.50 2.50 1.50 20.00 100.00 0.0346 0.00 1.7791 circular 1.50 2.14 1.50 20.00 100.00 0.0346 0.00 2.11
Model number hole type
434
Unstiffened element in uniaxial compression, offset holes hhole h Lhole S L t δhole fcrl
in. in. in. in. in. in. in. ksi1 slotted 1.50 15.00 4.00 20.00 100.00 0.0346 0.00 0.062 slotted 1.50 15.00 4.00 20.00 100.00 0.0346 -0.94 0.063 slotted 1.50 15.00 4.00 20.00 100.00 0.0346 -1.88 0.064 slotted 1.50 15.00 4.00 20.00 100.00 0.0346 -2.81 0.065 slotted 1.50 15.00 4.00 20.00 100.00 0.0346 -3.75 0.066 slotted 1.50 15.00 4.00 20.00 100.00 0.0346 -4.69 0.067 slotted 1.50 15.00 4.00 20.00 100.00 0.0346 -5.63 0.068 slotted 1.50 7.50 4.00 20.00 100.00 0.0346 0.00 0.239 slotted 1.50 7.50 4.00 20.00 100.00 0.0346 -0.47 0.23
10 slotted 1.50 7.50 4.00 20.00 100.00 0.0346 -0.94 0.2411 slotted 1.50 7.50 4.00 20.00 100.00 0.0346 -1.41 0.2412 slotted 1.50 7.50 4.00 20.00 100.00 0.0346 -1.88 0.2413 slotted 1.50 7.50 4.00 20.00 100.00 0.0346 -2.34 0.2514 slotted 1.50 5.00 4.00 20.00 100.00 0.0346 0.00 0.4915 slotted 1.50 5.00 4.00 20.00 100.00 0.0346 -0.31 0.5016 slotted 1.50 5.00 4.00 20.00 100.00 0.0346 -0.63 0.5117 slotted 1.50 5.00 4.00 20.00 100.00 0.0346 -0.94 0.5318 slotted 1.50 5.00 4.00 20.00 100.00 0.0346 -1.25 0.5519 slotted 1.50 3.75 4.00 20.00 100.00 0.0346 -0.07 0.8020 slotted 1.50 3.75 4.00 20.00 100.00 0.0346 -0.14 0.8121 slotted 1.50 3.75 4.00 20.00 100.00 0.0346 -0.20 0.8222 slotted 1.50 3.75 4.00 20.00 100.00 0.0346 -0.27 0.8323 slotted 1.50 3.75 4.00 20.00 100.00 0.0346 -0.34 0.8424 slotted 1.50 3.75 4.00 20.00 100.00 0.0346 -0.41 0.8625 slotted 1.50 3.75 4.00 20.00 100.00 0.0346 -0.47 0.8726 slotted 1.50 3.00 4.00 20.00 100.00 0.0346 -0.05 1.0527 slotted 1.50 3.00 4.00 20.00 100.00 0.0346 -0.09 1.0728 slotted 1.50 3.00 4.00 20.00 100.00 0.0346 -0.14 1.0929 slotted 1.50 3.00 4.00 20.00 100.00 0.0346 -0.18 1.1230 slotted 1.50 3.00 4.00 20.00 100.00 0.0346 -0.23 1.1431 slotted 1.50 3.00 4.00 20.00 100.00 0.0346 -0.27 1.1732 slotted 1.50 3.00 4.00 20.00 100.00 0.0346 -0.32 1.2033 slotted 1.50 2.50 4.00 20.00 100.00 0.0346 -0.03 1.1834 slotted 1.50 2.50 4.00 20.00 100.00 0.0346 -0.06 1.2135 slotted 1.50 2.50 4.00 20.00 100.00 0.0346 -0.09 1.2436 slotted 1.50 2.50 4.00 20.00 100.00 0.0346 -0.12 1.2737 slotted 1.50 2.50 4.00 20.00 100.00 0.0346 -0.15 1.3138 slotted 1.50 2.50 4.00 20.00 100.00 0.0346 -0.18 1.3539 slotted 1.50 2.50 4.00 20.00 100.00 0.0346 -0.21 1.3940 slotted 1.50 2.14 4.00 20.00 100.00 0.0346 -0.02 1.1641 slotted 1.50 2.14 4.00 20.00 100.00 0.0346 -0.04 1.1942 slotted 1.50 2.14 4.00 20.00 100.00 0.0346 -0.06 1.2243 slotted 1.50 2.14 4.00 20.00 100.00 0.0346 -0.08 1.2644 slotted 1.50 2.14 4.00 20.00 100.00 0.0346 -0.10 1.2945 slotted 1.50 2.14 4.00 20.00 100.00 0.0346 -0.12 1.3446 slotted 1.50 2.14 4.00 20.00 100.00 0.0346 -0.14 1.3847 slotted 1.50 15.00 4.00 20.00 100.00 0.0346 0.00 0.0648 slotted 1.50 15.00 4.00 20.00 100.00 0.0346 0.94 0.0649 slotted 1.50 15.00 4.00 20.00 100.00 0.0346 1.88 0.0650 slotted 1.50 15.00 4.00 20.00 100.00 0.0346 2.81 0.0651 slotted 1.50 15.00 4.00 20.00 100.00 0.0346 3.75 0.0652 slotted 1.50 15.00 4.00 20.00 100.00 0.0346 4.69 0.0653 slotted 1.50 15.00 4.00 20.00 100.00 0.0346 5.63 0.0654 slotted 1.50 7.50 4.00 20.00 100.00 0.0346 0.00 0.2355 slotted 1.50 7.50 4.00 20.00 100.00 0.0346 0.47 0.2356 slotted 1.50 7.50 4.00 20.00 100.00 0.0346 0.94 0.2357 slotted 1.50 7.50 4.00 20.00 100.00 0.0346 1.41 0.2258 slotted 1.50 7.50 4.00 20.00 100.00 0.0346 1.88 0.2259 slotted 1.50 7.50 4.00 20.00 100.00 0.0346 2.34 0.2260 slotted 1.50 5.00 4.00 20.00 100.00 0.0346 0.00 0.4961 slotted 1.50 5.00 4.00 20.00 100.00 0.0346 0.31 0.4862 slotted 1.50 5.00 4.00 20.00 100.00 0.0346 0.63 0.4763 slotted 1.50 5.00 4.00 20.00 100.00 0.0346 0.94 0.4664 slotted 1.50 5.00 4.00 20.00 100.00 0.0346 1.25 0.4565 slotted 1.50 3.75 4.00 20.00 100.00 0.0346 0.07 0.7766 slotted 1.50 3.75 4.00 20.00 100.00 0.0346 0.14 0.7667 slotted 1.50 3.75 4.00 20.00 100.00 0.0346 0.20 0.7568 slotted 1.50 3.75 4.00 20.00 100.00 0.0346 0.27 0.7569 slotted 1.50 3.75 4.00 20.00 100.00 0.0346 0.34 0.7470 slotted 1.50 3.75 4.00 20.00 100.00 0.0346 0.41 0.7371 slotted 1.50 3.75 4.00 20.00 100.00 0.0346 0.47 0.7272 slotted 1.50 3.00 4.00 20.00 100.00 0.0346 0.05 1.0173 slotted 1.50 3.00 4.00 20.00 100.00 0.0346 0.09 0.9974 slotted 1.50 3.00 4.00 20.00 100.00 0.0346 0.14 0.9775 slotted 1.50 3.00 4.00 20.00 100.00 0.0346 0.18 0.9676 slotted 1.50 3.00 4.00 20.00 100.00 0.0346 0.23 0.9477 slotted 1.50 3.00 4.00 20.00 100.00 0.0346 0.27 0.9278 slotted 1.50 3.00 4.00 20.00 100.00 0.0346 0.32 0.9179 slotted 1.50 2.50 4.00 20.00 100.00 0.0346 0.03 1.1380 slotted 1.50 2.50 4.00 20.00 100.00 0.0346 0.06 1.1081 slotted 1.50 2.50 4.00 20.00 100.00 0.0346 0.09 1.0882 slotted 1.50 2.50 4.00 20.00 100.00 0.0346 0.12 1.0583 slotted 1.50 2.50 4.00 20.00 100.00 0.0346 0.15 1.0384 slotted 1.50 2.50 4.00 20.00 100.00 0.0346 0.18 1.0185 slotted 1.50 2.50 4.00 20.00 100.00 0.0346 0.21 0.9986 slotted 1.50 2.14 4.00 20.00 100.00 0.0346 0.02 1.1087 slotted 1.50 2.14 4.00 20.00 100.00 0.0346 0.04 1.0788 slotted 1.50 2.14 4.00 20.00 100.00 0.0346 0.06 1.0589 slotted 1.50 2.14 4.00 20.00 100.00 0.0346 0.08 1.0290 slotted 1.50 2.14 4.00 20.00 100.00 0.0346 0.10 0.9991 slotted 1.50 2.14 4.00 20.00 100.00 0.0346 0.12 0.9792 slotted 1.50 2.14 4.00 20.00 100.00 0.0346 0.14 0.95
Model number hole type
435
Appendix C 11BDerivation of elastic buckling coefficients for unstiffened elements
C.1 kA for an unstiffened element in compression The finite strip method is employed with CUFSM (Schafer and Ádàny 2006) to
calculate the plate buckling coefficient for an unstiffened strip in compression, kA, as a
function of unstiffened strip aspect ratio (Lhole/hA) and the compressive stress ratio (ψA).
The unstiffened element model setup in CUFSM is provided in Figure 1833HC.1. The results
from the CUFSM parameter study, where ψA is varied from 0 to 1, are presented in
Figure 1834HC.2.
hA
Simply supported
f2
f1
1
2
ff
A =ψ
A
A
Section A-AUnstiffened Element
Lhole
free
Figure C.1 CUFSM finite strip modeling definition for an unstiffened element in compression
The fminsearch function in Matlab (Mathworks 2007) is used to determine the variables
α1 through α5 in the general equation form:
436
5
43
21
34.0578.0
α
αψα
ψααψ
⎟⎟⎠
⎞⎜⎜⎝
⎛++
−+
+=
A
holeA
A
AA
hL
k .
The variables are chosen to minimize the sum of the squared errors between the CUFSM
results in Figure 1835HC.2 and the fitted curve. The curve fitting results in the equation:
2
035.0024.0
76.170.234.0
578.0
⎟⎟⎠
⎞⎜⎜⎝
⎛++
−+
+=
A
holeA
A
AA
hL
k
ψ
ψψ
0 1 2 3 4 5 6 7 8 9 100
1
2
3
4
5
6
7
8
9
10
Lhole/hA
plat
e bu
cklin
g co
eff.,
kA
Figure C.2 The plate buckling coefficient kA for an unstiffened element in compression (the multiple curves represent 0≤ψA≤1 with a step of 0.1, 11 curves total)
The mean and standard deviation of the ABAQUS to predicted ratio when
0.1≤Lhole/yA≤2 is 1.14 and 0.61 respectively. As shown in Figure 1836HC.3, the accuracy of the
prediction is often conservative within this aspect ratio range. For 2<Lhole/yA≤10 the mean
and standard deviation of the ABAQUS to predicted ratio are 0.99 and 0.02 respectively.
437
0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 20
1
2
3
4
5
6
7
8
9
10
Lhole/hA
plat
e bu
cklin
g co
eff.,
kA
Fitted curve is conservative predictor of kA
Figure C.3 The fitted curve for kA is a conservative predictor when Lhole/yA≤2
C.2 kB for an unstiffened element with compression on the free edge
The finite strip method is employed with CUFSM (Schafer and Ádàny 2006) to
calculate the plate buckling coefficient for an unstiffened element with compression on
the free edge and tension on the simply supported edge, kB, as a function of unstiffened
strip aspect ratio (Lhole/hB) and the compressive stress ratio (ψB). The unstiffened element
model setup in CUFSM is provided in Figure 1837HC.4.
438
hA
Simply supported
f2
f1
1
2
ff
B =ψ
A
A
Section A-AUnstiffened Element
Lhole
free
Figure C.4 CUFSM finite strip modeling definition for an unstiffened element with compression on the free
edge, tension on the simply‐supported edge
The results from the CUFSM parameter study, where ψB is varied from 0 to 10, are
presented in Figure 1838HC.5 1839HC.5. As the portion of the plate that is in tension increases (i.e.,
ψB increases), the buckling mode switches from one buckled half‐wave to multiple half‐
waves.
0 5 10 150
5
10
15
20
25
30
35
40
Lhole/hB
plat
e bu
cklin
g co
eff.,
kB
Buckles in one half-wavelength (ψΒ=0 shown)
Buckles in several half-wavelength (ψΒ=10 shown)
Figure C.5 Variation in plate buckling coefficient kB for an unstiffened element with ψB ranging from 0 to 10
A polynomial curve is fit to the minimum kB as shown in Figure 1840HC.6:
573.0100.0340.0 2 ++= BBBk ψψ .
439
The mean and standard deviation of the CUFSM minima to fitted curve prediction ratio
are 1.03 and 0.11 respectively.
0 1 2 3 4 5 6 7 8 9 100
5
10
15
20
25
30
35
40
ψB
plat
e bu
cklin
g co
eff.,
kB
CUFSMFitted curve
Figure C.6 Curve fit to minimum kB for ψB ranging from 0 to 10
This approximation is accurate when Lhole/yB >5 but does not capture the boost in kB when
Lhole/hB is small. Since Lhole/hB may often be less than 1 when considering common plate and
hole sizes, it is important to have a viable estimate of kB to avoid overly conservative
predictions. A family of curves is fit to the CUFSM predictions for the case when
0.25≤Lhole/yB≤2 and where ψB is varied from 0 to 10, resulting in the following equation:
14.020.0
49.06.138.0
1.03.0
28.1
+⎟⎟⎠
⎞⎜⎜⎝
⎛+−
+⎟⎟⎠
⎞⎜⎜⎝
⎛+
=
hole
BB
hole
BB
B
Lh
Lh
k
ψ
ψ, 100,20 ≤≤≤≤ B
B
hole
hL ψ
440
The equation provides an accurate representation of kB as demonstrated in Figure 1841HC.7.
The mean and standard deviation of the CUFSM to predicted ratio are 0.01 and 0.03
respectively.
0 0.5 1 1.5 2 2.50
10
20
30
40
50
60
Lhole/hB
k B
Figure C.7 Family of curves used to simulate boost in kB when Lhole/hB≤2, ψB ranges from 0 to 10
441
Appendix D 12BElastic buckling prediction method of cross-sectional elements with holes
B1 Critical Elastic Buckling Stress of Elements with Perforations
B1.1 Uniformly Compressed Stiffened Element
For uniformly compressed stiffened elements with uniformly spaced perforations satisfying the limits
5.1hS
≥ and 2L
S
hole
≥ ,
the critical elastic buckling stress, fcrl, is [ ]crhcrcr f,fminf =l
. (Eq. B1.1-1)
The critical elastic buckling stress, fcr, without the influence of perforations is
( )
2
2
2
cr ht
112Ekf ⎟
⎠⎞
⎜⎝⎛
−=
μπ , (Eq. B1.1-2)
where k=4 for a stiffened element with L/h>4.
The critical elastic buckling stress, fcrh, with the influence of perforations is
( )hh1ff holenet,crhcrh −= , (Eq. B1.1-3)
where the critical elastic buckling stress, fcrh,net, at the location of a perforation is
[ ]crBcrAnet,crh f,fminf = . (Eq. B1.1-4)
The critical elastic buckling stress, fcri,of unstiffened strip i is
( )
2
i2
2
icri ht
112Ekf ⎟⎟
⎠
⎞⎜⎜⎝
⎛−
=ν
π and i = A or B, (Eq. B1.1-5)
where
1hL ihole ≥ , ( ) 6.0hL
2.0425.0k 95.0ihole
i−
+= (Eq. B1.1-6)
442
1hL ihole < , 925.0k i = , and i = A or B. (Eq. B1.1-7)
B1.2 Stiffened Element Under Stress Gradient
For stiffened elements under a stress gradient with uniformly spaced perforations satisfying the limits
5.1hS
≥ and 2L
S
hole
≥ ,
the critical elastic buckling stress, fcrl, is [ ]crhcrcr f,fminf =l
. (Eq. B1.2-1)
The critical elastic buckling stress, fcr, without the influence of perforations is
( )
2
2
2
cr ht
112Ekf ⎟
⎠⎞
⎜⎝⎛
−=
μπ , (Eq. B1.2-2)
where
( ) ( )ψψ ++++= 12124k 3 (Eq. B1.2-3)
and
( ) YYhff 12 −==ψ . (Eq. B1.2-4)
The critical elastic buckling stress, fcrh, with the influence of perforations is
for hA+hhole ≥ Y, ( )Yh1ff A
Anet,crhcrh ψ+= , and (Eq. B1.2-5)
for hA+hhole < Y, ⎥⎦
⎤⎢⎣
⎡⎟⎠⎞
⎜⎝⎛ −−=
Yh2
Yh1ff hole
Ahole
net,crhcrh ψ , (Eq. B1.2-6)
where
Y
hY AA
−=ψ . Eq. (B1.2-7)
The critical elastic buckling stress, fcrh,net, at the location of a perforation is
[ ]crBcrAnet,crh f,fminf = (Eq. B1.2-8)
Consideration of unstiffened strip “A” buckling is required only if hA<Y,
( )
2
A2
2
AcrA ht
112Ekf ⎟⎟
⎠
⎞⎜⎜⎝
⎛−
=ν
π (Eq. B1.2-9)
where
443
( )2
AholeA
A
AA hL035.0024.0
76.170.234.0
578.0k++
−+
+=
ψψ
ψ (Eq. B1.2-10)
Consideration of unstiffened strip “B” buckling is required only if hA+hhole<Y,
⎟⎟⎠
⎞⎜⎜⎝
⎛−−⎟⎟
⎠
⎞⎜⎜⎝
⎛−
=holeA
2
B2
2
BcrB hhYY
ht
)1(12Ekfν
π , (Eq. B1.2-12)
where
for Lhole/hB>2
573.0100.0340.0 2 ++= BBBk ψψ , (Eq. B1.2-12)
for Lhole/hB≤2
14.0Lh20.0
49.0Lh6.138.0
k 1.0
hole
B3.0B
2
hole
B8.1B
B
+⎟⎟⎠
⎞⎜⎜⎝
⎛+−
+⎟⎟⎠
⎞⎜⎜⎝
⎛+
=
ψ
ψ (Eq. B1.2-13)
and
holeA
B hhYYh−−
−=ψ , 100 B ≤≤ψ . (Eq. B1.2-14)
444
B1.3 Unstiffened Element Under Uniform Compression with Perforations
For uniformly compressed unstiffened elements with uniformly spaced perforations satisfying the limits
10h
L
A
hole ≤ , 10h
L
B
hole ≤ , 50.0h
h hole ≤ , and 2L
S
hole
≥ (Eq. B1.3-1)
the critical elastic buckling stress, fcrl, is [ ]crhcrcr f,fminf =l
. (Eq. B1.3-2)
The critical elastic buckling stress, fcr, without the influence of perforations is
( )
2
2
2
cr ht
112Ekf ⎟
⎠⎞
⎜⎝⎛
−=
μπ , (Eq. B1.3-3)
where k=0.425 for unstiffened elements with L/h>4.
The critical elastic buckling stress, fcrh, with the influence of perforations is
( ) ⎥
⎥⎦
⎤
⎢⎢⎣
⎡⎟⎠⎞
⎜⎝⎛ −⎟
⎠⎞
⎜⎝⎛
−=
hh
1f,ht
112Ekminf hole
crA
2
2
2
crh νπ , (Eq. B1.3-4)
where
⎟⎟⎠
⎞⎜⎜⎝
⎛−=
A
hole
hL
062.01425.0k (Eq. B1.3-5)
and fcrA is calculated with Eq. B1.1-5.
445
Appendix E 13BDerivation of global critical elastic buckling load for a column with holes
This derivation develops the equation for the flexural critical elastic buckling load of
a column with two holes spaced symmetrically about the longitudinal midline of a
column. The conclusions of this derivation are used as the foundation for the “weighted
properties” approach for approximating Pcre for columns and beams with holes as
described in Section 1842H4.2.7.3.1.1. INH is the moment of inertia of the column cross section
away from the hole and IH is the moment of inertia at the hole.
l1 l2
l3l4
LINH
IH
INH
IH
INHv
x
Figure E.1 Long column with two holes spaced symmetrically about the longitudinal midline.
A conservation of energy approach is employed in this derivation, and specifically
the Rayleigh‐Ritz Method. The derivation is founded on the fundamental principle
relating the strain energy and potential energy of the column, U and W respectively:
446
( ) 0=−=Π WUδδ
where the column strain energy U is:
dxdx
vdEIU ∫ ⎟⎟⎠
⎞⎜⎜⎝
⎛=
2
2
2
21
and the column potential energy is
dxdxdvPW
2
21
∫ ⎟⎠⎞
⎜⎝⎛= .
Applying the Raleigh‐Ritz method, we assume a shape function in the deformed
(buckled) configuration of the column:
( )Lxxv πα sin=
The derivatives of this function are calculated:
Lx
Ldxdv ππα cos=
Lx
Ldxvd ππα sin2
2
2
2
−=
and then substituted into the equations for U and W which are written along the length
of the column as:
dxLx
LEIdx
Lx
LEI
dxLx
LEIdx
Lx
LEIdx
Lx
LEIU
LHH
NHHNH
⎟⎠⎞
⎜⎝⎛+⎟
⎠⎞
⎜⎝⎛+
⎟⎠⎞
⎜⎝⎛+⎟
⎠⎞
⎜⎝⎛+⎟
⎠⎞
⎜⎝⎛=
∫∫
∫∫∫ππαππα
ππαππαππα
24
422
4
42
24
422
4
422
0 4
42
sin2
sin2
sin2
sin2
sin2
4
4
3
3
2
2
1
1
l
l
l
l
l
l
l
l
and
LPdx
Lx
LPW
L
4cos
21 22
0
22
22 παππα=⎟
⎠⎞
⎜⎝⎛= ∫ .
447
The derivative of U is taken with respect to the arbitrary shape function amplitude α:
⎥⎦
⎤⎢⎣
⎡⎟⎠⎞
⎜⎝⎛+⎟
⎠⎞
⎜⎝⎛++
⎥⎦
⎤⎢⎣
⎡⎟⎠⎞
⎜⎝⎛+⎟
⎠⎞
⎜⎝⎛+⎟
⎠⎞
⎜⎝⎛=
∫∫
∫∫∫
dxLxdx
Lx
LEI
dxLxdx
Lxdx
Lx
LEI
ddU
H
LNH
4
3
2
1
4
3
2
1
224
4
22
0
24
4
sinsin
sinsinsin
l
l
l
l
l
l
l
l
ππαπ
πππαπδ
The definite integrals inside the derivative are then solved resulting in:
⎥⎦
⎤⎢⎣
⎡⎟⎠⎞
⎜⎝⎛ +−+−++
⎥⎦
⎤⎢⎣
⎡⎟⎠⎞
⎜⎝⎛ ++−−+⎟
⎠⎞
⎜⎝⎛ −+−+=
LLLLLL
LEI
LLLLLL
LEI
ddU
HH
NH
34124
4
423142314
4
2sin2sin2sin2sin42
2sin2sin2sin2sin422222
llll
llllllll
πππππ
απ
πππππ
απδ
The length terms from the integration sum to the length of column without a hole, LNH,
and the length of column with a hole, LH. When the holes are symmetric about the
longitudinal midline of the column, the trigonometric terms cancel as shown in Figure
E.2 and the equation above simplifies to:
22 4
4
4
4HNHNHNH L
LEIL
LEI
ddU απαπ
α+=
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1-1.5
-1
-0.5
0
0.5
1
1.5
x/L
sin(
2 πx/
L)
H trig termNH trig term
Trig terms cancel for holes symmetric about longitudinal midline
Figure E.2 Trigonometric terms in energy solution cancel when holes are symmetric about longitudinal midline of column.
448
The solution for the potential energy of the column is not dependent upon the moment
of inertia and therefore the derivative can be solved directly as:
LP
ddW
2
2παα
=
The load P that minimizes the variation in energy is the critical elastic buckling load, Pcre:
( ) 0=−=−=Παα
δδddW
ddUWU
Equating the variational energy terms:
0222
2
4
4
4
4
=−+L
PLL
EILL
EI HHNHNH πααπαπ
results in a solution for Pcre where the moment of inertia is a weighted average of the net
and gross cross‐section of the column.
⎟⎠⎞
⎜⎝⎛ +
=L
LILIL
EP HHNHNHcre 2
2π
Pcre for a column with the general case of i=1..n holes can be approximated as:
( ) ( )⎟⎠⎞
⎜⎝⎛ +++
=L
TLITLIL
EP HHHNHNHNHcre 2
2π ,
where
⎟⎟⎠
⎞⎜⎜⎝
⎛⎟⎟⎠
⎞⎜⎜⎝
⎛= ∑
= LL
LLLT ihole
n
i
icNH
,
1
, sin2
cos2
πππ
, ⎟⎟⎠
⎞⎜⎜⎝
⎛⎟⎟⎠
⎞⎜⎜⎝
⎛−= ∑
= LL
LLLT ihole
n
i
icH
,
1
, sin2
cos2
πππ
.
Lc,i is the distance from the end of the column to the centerline of hole i and Lhole,i is the
length of hole i.
449
Appendix F 14BColumn experiment results
450
Column Specimen 362‐1‐24‐NH
0 0.05 0.1 0.15 0.2
-14
-12
-10
-8
-6
-4
-2
0
Column axial displacement (inches)
Col
umn
axia
l load
(kip
s)
Ptest=10.48 kips
362-1-24-NH
0 0.05 0.1 0.15 0.2-1.5
-1
-0.5
0
0.5
1
1.5
Column axial displacement (inches)
Flan
ge d
ispl
acem
ent (
inch
es)
362-1-24-NH
WE
Notes: Loaded N to S instead of S to N. Adjusted all geometry measurements. Lips rotated and not touching bottom platen after peak load.
451
Column Specimen 362‐2‐24‐NH
0 0.05 0.1 0.15 0.2
-14
-12
-10
-8
-6
-4
-2
0
Column axial displacement (inches)
Col
umn
axia
l load
(kip
s)
Ptest=10.51 kips
362-2-24-NH
0 0.05 0.1 0.15 0.2-1.5
-1
-0.5
0
0.5
1
1.5
Column axial displacement (inches)
Flan
ge d
ispl
acem
ent (
inch
es)
362-2-24-NH
WE
Notes:
452
Column Specimen 362‐3‐24‐NH
0 0.05 0.1 0.15 0.2
-14
-12
-10
-8
-6
-4
-2
0
Column axial displacement (inches)
Col
umn
axia
l load
(kip
s)
Ptest=10.15 kips
362-3-24-NH
0 0.05 0.1 0.15 0.2-1.5
-1
-0.5
0
0.5
1
1.5
Column axial displacement (inches)
Flan
ge d
ispl
acem
ent (
inch
es)
362-3-24-NH
WE
Notes: Bottom lips rotated at 7 kips post-peak and are not bearing on platen.
453
Column Specimen 362‐1‐24‐H
0 0.05 0.1 0.15 0.2
-14
-12
-10
-8
-6
-4
-2
0
Column axial displacement (inches)
Col
umn
axia
l load
(kip
s)
Ptest=10.01 kips
362-1-24-H
0 0.05 0.1 0.15 0.2-1.5
-1
-0.5
0
0.5
1
1.5
Column axial displacement (inches)
Flan
ge d
ispl
acem
ent (
inch
es)362-1-24-H
WE
Notes: Visible buckling of web on either side of hole at 7 kips.
454
Column Specimen 362‐2‐24‐H
0 0.05 0.1 0.15 0.2
-14
-12
-10
-8
-6
-4
-2
0
Column axial displacement (inches)
Col
umn
axia
l load
(kip
s)
Ptest=10.38 kips
362-2-24-H
0 0.05 0.1 0.15 0.2-1.5
-1
-0.5
0
0.5
1
1.5
Column axial displacement (inches)
Flan
ge d
ispl
acem
ent (
inch
es)
362-2-24-H
WE
Notes: Visible buckling of web on either side of hole at 5 kips.
Local buckling at hole (unstiffened strip)
(a) P=0 kips (b) P=10.4 kips(peak load)
(c) P=7.0 kips
455
Column Specimen 362‐3‐24‐H
0 0.05 0.1 0.15 0.2
-14
-12
-10
-8
-6
-4
-2
0
Column axial displacement (inches)
Col
umn
axia
l load
(kip
s)
Ptest=9.94 kips
362-3-24-H
0 0.05 0.1 0.15 0.2-1.5
-1
-0.5
0
0.5
1
1.5
Column axial displacement (inches)
Flan
ge d
ispl
acem
ent (
inch
es)
362-3-24-H
WE
Notes: Accidental preload to 3 kips when adjusting specimen for test.
456
Column Specimen 362‐1‐48‐NH
0 0.05 0.1 0.15 0.2
-14
-12
-10
-8
-6
-4
-2
0
Column axial displacement (inches)
Col
umn
axia
l load
(kip
s)
Ptest=9.09 kips
362-1-48-NH
0 0.05 0.1 0.15 0.2-1.5
-1
-0.5
0
0.5
1
1.5
Column axial displacement (inches)
Flan
ge d
ispl
acem
ent (
inch
es)
362-1-48-NH
WE
Notes: Good end conditions – no visible gaps. 9 kips – a metallic noise – yielding of west flange and increase in local wavelengths. Column failed by global-torsional collapse.
457
Column Specimen 362‐2‐48‐NH
Peak Load
0 0.05 0.1 0.15 0.2
-14
-12
-10
-8
-6
-4
-2
0
Column axial displacement (inches)
Col
umn
axia
l load
(kip
s)
Ptest=9.49 kips
362-2-48-NH
0 0.05 0.1 0.15 0.2-1.5
-1
-0.5
0
0.5
1
1.5
Column axial displacement (inches)
Flan
ge d
ispl
acem
ent (
inch
es)
362-2-48-NH
WE
Notes: Tight end conditions at 1.5 kips. Local buckling at 6.5 kips . No sounds for this test. Column failed by global-torsional collapse.
458
Column Specimen 362‐3‐48‐NH
Peak Load
0 0.05 0.1 0.15 0.2
-14
-12
-10
-8
-6
-4
-2
0
Column axial displacement (inches)
Col
umn
axia
l load
(kip
s)
Ptest=9.48 kips
362-3-48-NH
0 0.05 0.1 0.15 0.2-1.5
-1
-0.5
0
0.5
1
1.5
Column axial displacement (inches)
Flan
ge d
ispl
acem
ent (
inch
es)
362-3-48-NH
WE
Notes: Local buckling first observed at 6.5 kips. Local wavelengths lengthen at 8.5 kips. Yielding of flange lips at 9 kips (near peak). Column failed by global-torsional collapse.
459
Column Specimen 362‐1‐48‐H
Peak Load
0 0.05 0.1 0.15 0.2
-14
-12
-10
-8
-6
-4
-2
0
Column axial displacement (inches)
Col
umn
axia
l load
(kip
s)
Ptest=8.95 kips
362-1-48-H
0 0.05 0.1 0.15 0.2-1.5
-1
-0.5
0
0.5
1
1.5
Column axial displacement (inches)
Flan
ge d
ispl
acem
ent (
inch
es)
362-1-48-H
WE
Notes: No visible gaps and ends under 1 kip. Local buckling is visible at 7 kips. Local half-waves merge at 8.5 kips. Bulging of web at hole occurs near peak load. Column failed by global-torsional collapse.
460
Column Specimen 362‐2‐48‐H
Peak Load
0 0.05 0.1 0.15 0.2
-14
-12
-10
-8
-6
-4
-2
0
Column axial displacement (inches)
Col
umn
axia
l load
(kip
s)
Ptest=9.18 kips
362-2-48-H
0 0.05 0.1 0.15 0.2-1.5
-1
-0.5
0
0.5
1
1.5
Column axial displacement (inches)
Flan
ge d
ispl
acem
ent (
inch
es)
362-2-48-H
WE
Notes: End conditions tight at 4 kips. Local buckling visible at 6.5 kips. Distortional buckling seems to increase as load-displacement softens. East LVDT reaches limit of range as column starts to twist. Column failed by global-torsional collapse.
461
Column Specimen 362‐3‐48‐H
Peak Load
0 0.05 0.1 0.15 0.2
-14
-12
-10
-8
-6
-4
-2
0
Column axial displacement (inches)
Col
umn
axia
l load
(kip
s)
Ptest=9.37 kips
362-3-48-H
0 0.05 0.1 0.15 0.2-1.5
-1
-0.5
0
0.5
1
1.5
Column axial displacement (inches)
Flan
ge d
ispl
acem
ent (
inch
es)
362-3-48-H
WE
Notes: Local buckling visible at 6.5 kips. Column failed by global-torsional collapse.
462
Column Specimen 600‐1‐24‐NH
Peak Load
0 0.05 0.1 0.15 0.2
-14
-12
-10
-8
-6
-4
-2
0
Column axial displacement (inches)
Col
umn
axia
l load
(kip
s)
Ptest=11.93 kips
600-1-24-NH
0 0.05 0.1 0.15 0.2-1.5
-1
-0.5
0
0.5
1
1.5
Column axial displacement (inches)
Flan
ge d
ispl
acem
ent (
inch
es)
600-1-24-NH
WE
Notes: Local and distortional waves seem to stay separate. 8 kips (post-peak) – east flange buckles.
463
Column Specimen 600‐2‐24‐NH
Peak Load
0 0.05 0.1 0.15 0.2
-14
-12
-10
-8
-6
-4
-2
0
Column axial displacement (inches)
Col
umn
axia
l load
(kip
s)
Ptest=11.95 kips
600-2-24-NH
0 0.05 0.1 0.15 0.2-1.5
-1
-0.5
0
0.5
1
1.5
Column axial displacement (inches)
Flan
ge d
ispl
acem
ent (
inch
es)
600-2-24-NH
WE
Notes: Web has large curve when placing specimen on bottom platen. Visible gap between platen and specimen at top west web-flange corner - 5 kips. 10 kips (post-peak)- flanges buckle and lose contact with bottom platen.
464
Column Specimen 600‐3‐24‐NH
Peak Load
0 0.05 0.1 0.15 0.2
-14
-12
-10
-8
-6
-4
-2
0
Column axial displacement (inches)
Col
umn
axia
l load
(kip
s)
Ptest=12.24 kips
600-3-24-NH
0 0.05 0.1 0.15 0.2-1.5
-1
-0.5
0
0.5
1
1.5
Column axial displacement (inches)
Flan
ge d
ispl
acem
ent (
inch
es)
600-3-24-NH
WE
Notes: Specimen failed at bottom end condition, web rolled over and was not bearing on platen.
465
Column Specimen 600‐1‐24‐H
Peak Load
0 0.05 0.1 0.15 0.2
-14
-12
-10
-8
-6
-4
-2
0
Column axial displacement (inches)
Col
umn
axia
l load
(kip
s)
Ptest=12.14 kips
600-1-24-H
0 0.05 0.1 0.15 0.2-1.5
-1
-0.5
0
0.5
1
1.5
Column axial displacement (inches)
Flan
ge d
ispl
acem
ent (
inch
es)
600-1-24-H
WE
Notes: Slight gap at east top web-flange corner - 3 kips, gap is closed at 11 kips. East flange gives way at 11 kips with dip in load-disp. curve, may be related to above. Loud popping sound at 8 kips (post-peak) and large change in load-displ. slope.
466
Column Specimen 600‐2‐24‐H
Peak Load
0 0.05 0.1 0.15 0.2
-14
-12
-10
-8
-6
-4
-2
0
Column axial displacement (inches)
Col
umn
axia
l load
(kip
s)
Ptest=11.62 kips
600-2-24-H
0 0.05 0.1 0.15 0.2-1.5
-1
-0.5
0
0.5
1
1.5
Column axial displacement (inches)
Flan
ge d
ispl
acem
ent (
inch
es)
600-2-24-H
WE
Notes: Specimen failure mode similar to that of a no-hole specimen.
467
Column Specimen 600‐3‐24‐H
Peak Load
0 0.05 0.1 0.15 0.2
-14
-12
-10
-8
-6
-4
-2
0
Column axial displacement (inches)
Col
umn
axia
l load
(kip
s)
Ptest=11.79 kips
600-3-24-H
0 0.05 0.1 0.15 0.2-1.5
-1
-0.5
0
0.5
1
1.5
Column axial displacement (inches)
Flan
ge d
ispl
acem
ent (
inch
es)
600-3-24-H
WE
Notes: Good contact with platens.
468
Column Specimen 600‐1‐48‐NH
Peak Load
0 0.05 0.1 0.15 0.2
-14
-12
-10
-8
-6
-4
-2
0
Column axial displacement (inches)
Col
umn
axia
l load
(kip
s)
Ptest=11.15 kips
600-1-48-NH
0 0.05 0.1 0.15 0.2-1.5
-1
-0.5
0
0.5
1
1.5
Column axial displacement (inches)
Flan
ge d
ispl
acem
ent (
inch
es)
600-1-48-NH
WE
Notes: Local buckling first observed at 4.5 kips (11 half-waves). Distortional wave becomes prominent at 10 kips. Loud noises 1 minute apart – L waves turn to D waves at north, then south ends.
469
Column Specimen 600‐2‐48‐NH
Peak Load
0 0.05 0.1 0.15 0.2
-14
-12
-10
-8
-6
-4
-2
0
Column axial displacement (inches)
Col
umn
axia
l load
(kip
s)
Ptest=11.44 kips
600-2-48-NH
0 0.05 0.1 0.15 0.2-1.5
-1
-0.5
0
0.5
1
1.5
Column axial displacement (inches)
Flan
ge d
ispl
acem
ent (
inch
es)
600-2-48-NH
WE
Notes: Gap between platen and specimen at top east flange-web corner closes at 2 kips. Can see distortional shape developing at 4.5 kips. Local buckling visible at 5 kips. Two loud bangs (peak load, 10.5 kips post peak) – local web waves change to D waves. Flange distortion slows at 7 kips post-peak.
470
Column Specimen 600‐3‐48‐NH
Peak Load
0 0.05 0.1 0.15 0.2
-14
-12
-10
-8
-6
-4
-2
0
Column axial displacement (inches)
Col
umn
axia
l load
(kip
s)
Ptest=11.29 kips
600-3-48-NH
0 0.05 0.1 0.15 0.2-1.5
-1
-0.5
0
0.5
1
1.5
Column axial displacement (inches)
Flan
ge d
ispl
acem
ent (
inch
es)
600-3-48-NH
WE
Notes: Gap between platen and specimen at east top flange closes at 1 kip. Loud sound at peak load – L waves change to D waves in web.
471
Column Specimen 600‐1‐48‐H
Peak Load
0 0.05 0.1 0.15 0.2
-14
-12
-10
-8
-6
-4
-2
0
Column axial displacement (inches)
Col
umn
axia
l load
(kip
s)
Ptest=11.16 kips
600-1-48-H
0 0.05 0.1 0.15 0.2-1.5
-1
-0.5
0
0.5
1
1.5
Column axial displacement (inches)
Flan
ge d
ispl
acem
ent (
inch
es)
600-1-48-H
WE
Notes: Local buckling and DH mode visible at 5 kips. Loud noise at 9.5 kips – L waves changes to D wave in web.
472
Column Specimen 600‐2‐48‐H
Peak Load
0 0.05 0.1 0.15 0.2
-14
-12
-10
-8
-6
-4
-2
0
Column axial displacement (inches)
Col
umn
axia
l load
(kip
s)
Ptest=11.7 kips
600-2-48-H
0 0.05 0.1 0.15 0.2-1.5
-1
-0.5
0
0.5
1
1.5
Column axial displacement (inches)
Flan
ge d
ispl
acem
ent (
inch
es)
600-2-48-H
WE
Notes: Local buckling visible at 4 kips. D wave interrupted by large crease. L web waves change to D waves (9 kips post-peak, 7.5 kips post-peak)
473
Column Specimen 600‐3‐48‐H
Peak Load
0 0.05 0.1 0.15 0.2
-14
-12
-10
-8
-6
-4
-2
0
Column axial displacement (inches)
Col
umn
axia
l load
(kip
s)
Ptest=11.16 kips
600-3-48-H
0 0.05 0.1 0.15 0.2-1.5
-1
-0.5
0
0.5
1
1.5
Column axial displacement (inches)
Flan
ge d
ispl
acem
ent (
inch
es)
600-3-48-H
WE
Notes: Good platen bearing conditions. Loud noise at 7.5 kips post-peak. Yielding in the west flange first, then east flange.
474
Appendix G 15BResidual stresses– backstress for kinematic hardening implementation
Implementation of a kinematic hardening rule requires that the center of the yield
surface, in stress space, be known for any material which has been yielded prior to the
loading of interest. The coordinates of the center of the yield surface (Δσ1, Δσ2, Δσ3),
known as the backstress, cannot be directly calculated from the stresses derived herein
because work hardening was ignored in the residual stress derivations. However, the
plastic strains developed in the manufacturing process provide a means by which the
backstress may be approximated, as provided in this appendix.
The general equation for effective stress is defined as
( ) ( ) ( )213
232
2212
1 σσσσσσσ −+−+−=e . (G.1)
Given that the through‐thickness sheet stresses are zero (σ2=0), Eq. (G.1) reduces to
2331
21 σσσσσ +−=e (G.2)
Consider the contribution to the backstress that develops due to coiling. From Eq.
1843H(6.18) we know the plastic strain, εpcoiling. With εpcoiling and knowing the material stress‐
strain relation (i.e., 1844HFigure 6.23) the effective stress at that plastic strain, σeycoiling maybe
determined. Consistent with the residual stress derivation of Eq. 1845H(6.8), we assume ν=0.3
and
475
coilingcoiling31 νσσ = . (G.3)
Finally, substituting the preceding into Eq. (G.2) results in
123
+−=
νν
σσ
coilingeycoiling . (G.4)
Similarly for cold bending the corners, from Eq. 1846H(6.21) we know the plastic strain,
εpbend. With εpbend and knowing the material stress‐strain relation (i.e., 1847HFigure 6.23), we
determine the effective stress at that plastic strain, σeybend. Consistent with the residual
stress derivation of Eq. 1848H(6.12), we assume ν=0.5 and
bendbend13 νσσ = , (G.5)
and thus find
121
+−=
νν
σσ
bendeybend . (G.6)
The backstress is then determined as:
yieldbendcoiling σσσσ −+=Δ 111
02 =Δσ (G.7)
yieldbendcoiling σσσσ −+=Δ 333 ,
where σyield is the virgin yield stress of the steel. This estimate assumes that the changes
in material properties from coiling, uncoiling, and flattening and cold‐forming do not
influence one another.
476
Appendix H 16BExperiment true stress-strain curves
The average true plastic stress‐strain curves are provided here for each of the 24
column tests reported in 1849HChapter 5. For each specimen, three engineering stress‐strain
curves (west flange, east flange, and web) were averaged and then transformed into true
stresses and strains with the following equations:
)1()1ln(
ootrue
otrue
εσσεε
+=+=
εtrue and σtrue are the true stress and strain and εo and σo are the engineering stress and
strain in the above equations. The tables in this appendix provide just the plastic
component of the true strain since this is what is required in ABAQUS:
yieldtruep εεε −= , where Eyield
yield
σε =
The true stress‐strain curves presented here were modified prior to their
implementation in ABAQUS to ensure plasticity initiated at the yield stress and not
the proportional limit. Refer to Section 1850H7.2.1.4 for details on this modeling decision.
477
Specimen 362‐1‐24‐NH, 362‐2‐24‐NH, 362‐3‐24‐NH
0 0.02 0.04 0.06 0.08 0.1 0.120
10
20
30
40
50
60
70
80
90
100
Avg. Yield Stress=55.1 ksi
true plastic strain
true
stre
ss, k
si
true plastic strain, εp
true stress, σtrue
ksi0 33.05
0.001 46.100.002 51.880.007 60.300.012 64.890.017 68.370.027 73.970.037 78.120.047 81.270.057 83.830.067 86.16
478
Specimen 362‐1‐24‐H
0 0.02 0.04 0.06 0.08 0.1 0.120
10
20
30
40
50
60
70
80
90
100
Avg. Yield Stress=57.9 ksi
true plastic strain
true
stre
ss, k
si
true plastic strain, εp
true stress, σtrue
ksi0 34.2
0.001 50.00.002 56.10.007 64.40.012 68.30.017 72.00.027 78.60.037 82.50.047 86.20.057 88.70.067 91.00.077 92.90.087 94.60.097 96.20.107 97.5
479
Specimen 362‐2‐24‐H
0 0.02 0.04 0.06 0.08 0.1 0.120
10
20
30
40
50
60
70
80
90
100
Avg. Yield Stress=57.1 ksi
true plastic strain
true
stre
ss, k
si
true plastic strain, εp
true stress, σtrue
ksi0 27.7
0.001 45.10.002 53.30.007 62.80.012 67.10.017 71.20.027 76.70.037 81.30.047 84.70.057 87.40.067 89.70.077 91.60.087 93.40.097 94.80.107 96.2
480
Specimen 362‐3‐24‐H
0 0.02 0.04 0.06 0.08 0.1 0.120
10
20
30
40
50
60
70
80
90
100
Avg. Yield Stress=56 ksi
true plastic strain
true
stre
ss, k
si
true plastic strain, εp
true stress, σtrue
ksi0 31.7
0.001 47.80.002 53.60.007 61.10.012 64.70.017 68.40.027 74.80.037 78.60.047 82.20.057 84.60.067 86.90.077 88.80.087 90.40.097 91.90.107 93.1
481
Specimen 362‐1‐48‐NH
0 0.02 0.04 0.06 0.08 0.1 0.120
10
20
30
40
50
60
70
80
90
100
Avg. Yield Stress=59.7 ksi
true plastic strain
true
stre
ss, k
si
true plastic strain, εp
true stress, σtrue
ksi0 37.8
0.001 51.60.002 56.60.007 64.30.012 68.20.017 72.00.027 78.10.037 82.10.047 85.90.057 88.30.067 90.80.077 92.50.087 94.30.097 95.70.107 97.1
482
Specimen 362‐2‐48‐NH
0 0.02 0.04 0.06 0.08 0.1 0.120
10
20
30
40
50
60
70
80
90
100
Avg. Yield Stress=59.2 ksi
true plastic strain
true
stre
ss, k
si
true plastic strain, εp
true stress, σtrue
ksi0 35.0
0.001 50.90.002 56.50.007 64.30.012 68.30.017 72.10.027 78.20.037 82.30.047 86.10.057 88.50.067 90.90.077 92.80.087 94.50.097 96.00.107 97.3
483
Specimen 362‐3‐48‐NH
0 0.02 0.04 0.06 0.08 0.1 0.120
10
20
30
40
50
60
70
80
90
100
Avg. Yield Stress=59 ksi
true plastic strain
true
stre
ss, k
si
true plastic strain, εp
true stress, σtrue
ksi0 34
0.001 500.002 560.007 640.012 680.017 720.027 780.037 820.047 860.057 880.067 910.077 920.087 940.097 960.107 97
484
Specimen 362‐1‐48‐H
0 0.02 0.04 0.06 0.08 0.1 0.120
10
20
30
40
50
60
70
80
90
100
Avg. Yield Stress=58.6 ksi
true plastic strain
true
stre
ss, k
si
true plastic strain, εp
true stress, σtrue
ksi0 33.2
0.001 50.10.002 56.10.007 64.20.012 68.20.017 72.00.027 78.00.037 82.20.047 85.70.057 88.40.067 90.70.077 92.70.087 94.40.097 95.80.107 97.2
485
Specimen 362‐2‐48‐H
0 0.02 0.04 0.06 0.08 0.1 0.120
10
20
30
40
50
60
70
80
90
100
Avg. Yield Stress=59.7 ksi
true plastic strain
true
stre
ss, k
si
true plastic strain, εp
true stress, σtrue
ksi0 37.3
0.001 51.50.002 56.70.007 64.20.012 68.20.017 72.00.027 78.10.037 82.40.047 85.80.057 88.50.067 90.80.077 92.70.087 94.40.097 95.90.107 97.2
486
Specimen 362‐3‐48‐H
0 0.02 0.04 0.06 0.08 0.1 0.120
10
20
30
40
50
60
70
80
90
100
Avg. Yield Stress=58.3 ksi
true plastic strain
true
stre
ss, k
si
true plastic strain, εp
true stress, σtrue
ksi0 37.5
0.001 50.60.002 55.60.007 62.90.012 67.10.017 70.90.027 76.50.037 80.80.047 84.10.057 86.80.067 89.10.077 90.90.087 92.60.097 94.10.107 95.4
487
Specimen 600‐1‐24‐NH, 600‐2‐24‐NH, 600‐3‐24‐NH
0 0.02 0.04 0.06 0.08 0.1 0.120
10
20
30
40
50
60
70
80
90
100
Avg. Yield Stress=58.7 ksi
true plastic strain
true
stre
ss, k
si
true plastic strain, εp
true stress, σtrue
ksi0 37.5
0.001 54.70.002 58.30.007 60.00.012 61.50.017 64.00.027 70.20.037 74.40.047 77.50.057 80.00.067 81.90.077 83.50.087 84.90.097 86.10.107 87.2
488
Specimen 600‐1‐24‐H
0 0.02 0.04 0.06 0.08 0.1 0.120
10
20
30
40
50
60
70
80
90
100
Avg. Yield Stress=61.9 ksi
true plastic strain
true
stre
ss, k
si
true plastic strain, εp
true stress, σtrue
ksi0 35.0
0.001 58.90.002 61.40.007 62.30.012 62.90.017 63.60.027 69.40.037 74.40.047 77.80.057 80.50.067 82.60.077 84.40.087 85.90.097 87.20.107 88.4
489
Specimen 600‐2‐24‐H
0 0.02 0.04 0.06 0.08 0.1 0.120
10
20
30
40
50
60
70
80
90
100
Avg. Yield Stress=58.4 ksi
true plastic strain
true
stre
ss, k
si
true plastic strain, εp
true stress, σtrue
ksi0 33.0
0.001 54.00.002 57.20.007 58.90.012 61.50.017 64.40.027 70.50.037 74.80.047 78.00.057 80.40.067 82.40.077 84.10.087 85.50.097 86.70.107 87.9
490
Specimen 600‐3‐24‐H
0 0.02 0.04 0.06 0.08 0.1 0.120
10
20
30
40
50
60
70
80
90
100
Avg. Yield Stress=60.1 ksi
true plastic strain
true
stre
ss, k
si
true plastic strain, εp
true stress, σtrue
ksi0 34.9
0.001 57.50.002 60.10.007 60.80.012 61.80.017 63.60.027 69.80.037 74.10.047 77.60.057 79.90.067 82.00.077 83.70.087 85.10.097 86.50.107 87.6
491
Specimen 600‐1‐48‐NH
0 0.02 0.04 0.06 0.08 0.1 0.120
10
20
30
40
50
60
70
80
90
100
Avg. Yield Stress=60.1 ksi
true plastic strain
true
stre
ss, k
si
true plastic strain, εp
true stress, σtrue
ksi0 44.4
0.001 58.90.002 60.30.007 61.00.012 62.00.017 64.80.027 70.60.037 75.00.047 78.30.057 80.80.067 82.80.077 84.50.087 86.00.097 87.30.107 88.5
492
Specimen 600‐2‐48‐NH
0 0.02 0.04 0.06 0.08 0.1 0.120
10
20
30
40
50
60
70
80
90
100
Avg. Yield Stress=63.4 ksi
true plastic strain
true
stre
ss, k
si
true plastic strain, εp
true stress, σtrue
ksi0 41.9
0.001 62.10.002 63.10.007 63.60.012 64.00.017 64.50.027 69.80.037 75.00.047 78.50.057 81.20.067 83.50.077 85.30.087 87.00.097 88.40.107 89.6
493
Specimen 600‐3‐48‐NH
0 0.02 0.04 0.06 0.08 0.1 0.120
10
20
30
40
50
60
70
80
90
100
Avg. Yield Stress=61.2 ksi
true plastic strain
true
stre
ss, k
si
true plastic strain, εp
true stress, σtrue
ksi0 36.0
0.001 55.60.002 60.00.007 61.70.012 62.00.017 62.50.027 68.50.037 73.40.047 76.90.057 79.40.067 81.60.077 83.30.087 84.90.097 86.20.107 87.4
494
Specimen 600‐1‐48‐H
0 0.02 0.04 0.06 0.08 0.1 0.120
10
20
30
40
50
60
70
80
90
100
Avg. Yield Stress=61.4 ksi
true plastic strain
true
stre
ss, k
si
true plastic strain, εp
true stress, σtrue
ksi0 37.5
0.001 57.70.002 60.90.007 61.90.012 62.40.017 64.30.027 70.40.037 75.00.047 78.40.057 81.00.067 83.10.077 84.80.087 86.30.097 87.70.107 88.9
495
Specimen 600‐2‐48‐H
0 0.02 0.04 0.06 0.08 0.1 0.120
10
20
30
40
50
60
70
80
90
100
Avg. Yield Stress=62 ksi
true plastic strain
true
stre
ss, k
si
true plastic strain, εp
true stress, σtrue
ksi0 37.4
0.001 57.60.002 61.60.007 62.60.012 63.30.017 64.10.027 69.40.037 74.20.047 77.80.057 80.60.067 82.90.077 84.80.087 86.40.097 87.80.107 89.1
496
Specimen 600‐3‐48‐H
0 0.02 0.04 0.06 0.08 0.1 0.120
10
20
30
40
50
60
70
80
90
100
Avg. Yield Stress=61.5 ksi
true plastic strain
true
stre
ss, k
si
true plastic strain, εp
true stress, σtrue
ksi0 32.4
0.001 57.10.002 61.30.007 62.20.012 62.60.017 64.50.027 70.40.037 75.20.047 78.60.057 81.10.067 83.20.077 84.90.087 86.50.097 87.80.107 89.0
497
Appendix I 17BColumn experiment nonlinear FE simulation results
Specimen 362‐1‐24‐NH
0 0.02 0.04 0.06 0.08 0.1 0.12 0.14 0.16 0.18 0.20
2
4
6
8
10
12
14
axial displacement, in.
axia
l loa
d, k
ips
ExperimentFE, 25% Imperfection CDFFE, 75% imperfection CDFFE, measured imperfections
498
Specimen 362‐2‐24‐NH
0 0.02 0.04 0.06 0.08 0.1 0.12 0.14 0.16 0.18 0.20
2
4
6
8
10
12
14
axial displacement, in.
axia
l loa
d, k
ips
ExperimentFE, 25% Imperfection CDFFE, 75% imperfection CDFFE, measured imperfections
499
Specimen 362‐3‐24‐NH
0 0.02 0.04 0.06 0.08 0.1 0.12 0.14 0.16 0.18 0.20
2
4
6
8
10
12
14
axial displacement, in.
axia
l loa
d, k
ips
ExperimentFE, 25% Imperfection CDFFE, 75% imperfection CDFFE, measured imperfections
500
Specimen 362‐1‐24‐H
0 0.02 0.04 0.06 0.08 0.1 0.12 0.14 0.16 0.18 0.20
2
4
6
8
10
12
14
axial displacement, in.
axia
l loa
d, k
ips
ExperimentFE, 25% Imperfection CDFFE, 75% imperfection CDFFE, measured imperfections
501
Specimen 362‐2‐24‐H
0 0.02 0.04 0.06 0.08 0.1 0.12 0.14 0.16 0.18 0.20
2
4
6
8
10
12
14
axial displacement, in.
axia
l loa
d, k
ips
ExperimentFE, 25% Imperfection CDFFE, 75% imperfection CDFFE, measured imperfections
502
Specimen 362‐3‐24‐H
0 0.02 0.04 0.06 0.08 0.1 0.12 0.14 0.16 0.18 0.20
2
4
6
8
10
12
14
axial displacement, in.
axia
l loa
d, k
ips
ExperimentFE, 25% Imperfection CDFFE, 75% imperfection CDFFE, measured imperfections
503
Specimen 362‐1‐48‐NH
0 0.02 0.04 0.06 0.08 0.1 0.12 0.14 0.16 0.18 0.20
2
4
6
8
10
12
14
axial displacement, in.
axia
l loa
d, k
ips
ExperimentFE, 25% Imperfection CDFFE, 75% imperfection CDFFE, measured imperfections
504
Specimen 362‐2‐48‐NH
0 0.02 0.04 0.06 0.08 0.1 0.12 0.14 0.16 0.18 0.20
2
4
6
8
10
12
14
axial displacement, in.
axia
l loa
d, k
ips
ExperimentFE, 25% Imperfection CDFFE, 75% imperfection CDFFE, measured imperfections
505
Specimen 362‐3‐48‐NH
0 0.02 0.04 0.06 0.08 0.1 0.12 0.14 0.16 0.18 0.20
2
4
6
8
10
12
14
axial displacement, in.
axia
l loa
d, k
ips
ExperimentFE, 25% Imperfection CDFFE, 75% imperfection CDFFE, measured imperfections
506
Specimen 362‐1‐48‐H
0 0.02 0.04 0.06 0.08 0.1 0.12 0.14 0.16 0.18 0.20
2
4
6
8
10
12
14
axial displacement, in.
axia
l loa
d, k
ips
ExperimentFE, 25% Imperfection CDFFE, 75% imperfection CDFFE, measured imperfections
507
Specimen 362‐2‐48‐H
0 0.02 0.04 0.06 0.08 0.1 0.12 0.14 0.16 0.18 0.20
2
4
6
8
10
12
14
axial displacement, in.
axia
l loa
d, k
ips
ExperimentFE, 25% Imperfection CDFFE, 75% imperfection CDFFE, measured imperfections
508
Specimen 362‐3‐48‐H
0 0.02 0.04 0.06 0.08 0.1 0.12 0.14 0.16 0.18 0.20
2
4
6
8
10
12
14
axial displacement, in.
axia
l loa
d, k
ips
ExperimentFE, 25% Imperfection CDFFE, 75% imperfection CDF
509
Specimen 600‐1‐24‐NH
0 0.02 0.04 0.06 0.08 0.1 0.12 0.14 0.16 0.18 0.20
2
4
6
8
10
12
14
axial displacement, in.
axia
l loa
d, k
ips
ExperimentFE, 25% Imperfection CDFFE, 75% imperfection CDFFE, measured imperfections
510
Specimen 600‐2‐24‐NH
0 0.02 0.04 0.06 0.08 0.1 0.12 0.14 0.16 0.18 0.20
2
4
6
8
10
12
14
axial displacement, in.
axia
l loa
d, k
ips
ExperimentFE, 25% Imperfection CDFFE, 75% imperfection CDFFE, measured imperfections
511
Specimen 600‐3‐24‐NH
0 0.02 0.04 0.06 0.08 0.1 0.12 0.14 0.16 0.18 0.20
2
4
6
8
10
12
14
axial displacement, in.
axia
l loa
d, k
ips
ExperimentFE, 25% Imperfection CDFFE, 75% imperfection CDFFE, measured imperfections
512
Specimen 600‐1‐24‐H
0 0.02 0.04 0.06 0.08 0.1 0.12 0.14 0.16 0.18 0.20
2
4
6
8
10
12
14
axial displacement, in.
axia
l loa
d, k
ips
ExperimentFE, 25% Imperfection CDFFE, 75% imperfection CDFFE, measured imperfections
513
Specimen 600‐2‐24‐H
0 0.02 0.04 0.06 0.08 0.1 0.12 0.14 0.16 0.18 0.20
2
4
6
8
10
12
14
axial displacement, in.
axia
l loa
d, k
ips
ExperimentFE, 25% Imperfection CDFFE, 75% imperfection CDFFE, measured imperfections
514
Specimen 600‐3‐24‐H
0 0.02 0.04 0.06 0.08 0.1 0.12 0.14 0.16 0.18 0.20
2
4
6
8
10
12
14
axial displacement, in.
axia
l loa
d, k
ips
ExperimentFE, 25% Imperfection CDFFE, 75% imperfection CDFFE, measured imperfections
515
Specimen 600‐1‐48‐NH
0 0.02 0.04 0.06 0.08 0.1 0.12 0.14 0.16 0.18 0.20
2
4
6
8
10
12
14
axial displacement, in.
axia
l loa
d, k
ips
ExperimentFE, 25% Imperfection CDFFE, 75% imperfection CDFFE, measured imperfections
516
Specimen 600‐2‐48‐NH
0 0.02 0.04 0.06 0.08 0.1 0.12 0.14 0.16 0.18 0.20
2
4
6
8
10
12
14
axial displacement, in.
axia
l loa
d, k
ips
ExperimentFE, 25% Imperfection CDFFE, 75% imperfection CDFFE, measured imperfections
517
Specimen 600‐3‐48‐NH
0 0.02 0.04 0.06 0.08 0.1 0.12 0.14 0.16 0.18 0.20
2
4
6
8
10
12
14
axial displacement, in.
axia
l loa
d, k
ips
ExperimentFE, 25% Imperfection CDFFE, 75% imperfection CDFFE, measured imperfections
518
Specimen 600‐1‐48‐H
0 0.02 0.04 0.06 0.08 0.1 0.12 0.14 0.16 0.18 0.20
2
4
6
8
10
12
14
axial displacement, in.
axia
l loa
d, k
ips
ExperimentFE, 25% Imperfection CDFFE, 75% imperfection CDF
519
Specimen 600‐2‐48‐H
0 0.02 0.04 0.06 0.08 0.1 0.12 0.14 0.16 0.18 0.20
2
4
6
8
10
12
14
axial displacement, in.
axia
l loa
d, k
ips
ExperimentFE, 25% Imperfection CDFFE, 75% imperfection CDFFE, measured imperfections
520
Specimen 600‐3‐48‐H
0 0.02 0.04 0.06 0.08 0.1 0.12 0.14 0.16 0.18 0.20
2
4
6
8
10
12
14
axial displacement, in.
axia
l loa
d, k
ips
ExperimentFE, 25% Imperfection CDFFE, 75% imperfection CDFFE, measured imperfections
521
Appendix J 18BContact simulation in ABAQUS
The friction‐bearing end conditions were chosen for the experimental program
described in 1851HChapter 5 because they allowed for convenient alignment and testing of the
column specimens. These boundary conditions were expected to behave as fixed‐fixed,
although during the test slipping of the cross‐section and lifting off of the specimens
were observed for some specimens in the post‐peak region of the load‐displacement
curve. A nonlinear FE study was performed in ABAQUS where the experiment friction‐
bearing boundary conditions were replicated using contact modeling (Moen and Schafer
2007a). These end conditions allowed deformation of the cross‐section at the bearing
ends under load (slipping) and lift off of the bearing ends. A master analytical rigid
surface was defined to represent the top and bottom platen as shown in Figure J.1 and
each surface was assigned a reference node. The rigid surfaces simulate fixed‐fixed
conditions by restraining the reference node degrees of freedom, and the specimen was
loaded by applying an imposed displacement to the bottom surface reference node. Top
and bottom node‐based slave surfaces were defined to simulate the bearing end of each
specimen. The tributary bearing area was defined at each node in the slave surface to
ensure that contact stresses were simulated accurately in ABAQUS.
522
1
2
3
45
6
ABAQUS Analytical Rigid Surface (Typ.)
Restrain rigid surface reference node in 2 to 6 directions
Restrain rigid surface reference node in 1 to 6 directions
Apply imposed displacement of surface in 1 direction
Assign friction contact behavior in ABAQUS between rigid surface and specimen
Figure J.1 Contact boundary condition as implemented in ABAQUS
A Coulomb friction model was enforced in ABAQUS between the master and slave
surfaces by defining a static and kinetic coefficient of friction, μs and μk, for steel‐on‐steel
contact. The assumed values for μs and μk were 0.7 and 0.6 in this study (Oden and
Martins 1985). Slip occurs in the model once the shear stress at the contact interface
exceeds μsfn, where fn is the normal contact stress at the bearing surface.
The locations of the rigid surfaces were defined to be in contact with the specimen
ends when the first step of the analysis began. This does not guarantee perfect contact in
a computational sense, and so the ABAQUS command ADJUST was used to zero the
contact surface and avoid numerical instabilities during the first analysis step. The
ADJUST command modifies the geometry of the specimen to close infinitesimal gaps,
but does not result in internal forces or moments in the specimen.
523
To evaluate the influence of the contact boundary conditions, the load‐displacement
response of specimen 600‐2‐24‐NH assuming contact boundary conditions was
evaluated against an FE simulation employing the fixed‐fixed boundary conditions
described in 1852HFigure 7.27. The results of the two simulations are almost identical until
well into the post‐peak range as shown in Figure J.2, demonstrating that the fixed‐fixed
boundary conditions are a viable approximation to the actual boundary conditions in an
FE simulation.
0 0.05 0.1 0.15 0.20
2
4
6
8
10
12
14
16
18
20
axial displacement, in.
colu
mn
axia
l loa
d, k
ips
Contact SurfaceFixed-Fixed
The failure response with friction-bearing boundary conditions is similar to that with fixed-fixed boundary conditions
Figure J.2. A comparison of ABAQUS nonlinear solutions considering fixed‐fixed and contact boundary
conditions for specimen 600‐2‐24‐NH
524
Appendix K 19BSimulated column experiments database
The table provided in this appendix summarizes the dimensions, elastic buckling loads,
and tested strengths of simulated column experiments described in Section 1853H8.1.11854H0. The
letters in the “Study type” column denote simulations considered in the DSM failure
mode studies: “D” for the distortional buckling failure study in Section 1855H8.1.2, “G” for the
global buckling failure study in Section 1856H8.1.3, and “L” for the local buckling failure study
in Section 1857H8.1.4.
The database contains columns with slotted holes or circular holes. To identify the hole
type in a specific column, use the following rule: columns with slotted holes always
have Lhole=4 in, and Lhole = hhole for columns with circular holes.
The following notation is employed to denote simulated strengths:
Ptest25 25% CDF local and distortional buckling, no global imperfectionsPtest75 75% CDF local and distortional buckling, no global imperfectionsPtest25+ 25% CDF local and distortional buckling, +L/2000 global imperfectionPtest75+ 75% CDF local and distortional buckling, +L/1000 global imperfectionPtest25- 25% CDF local and distortional buckling, -L/2000 global imperfectionPtest75- 75% CDF local and distortional buckling, -L/1000 global imperfection
525
ID # SSMA L Lhole hhole S # of Pyg Pynet Pynet/Pyg Pcrl Pcrd Pcre Ptest25 Ptest75 Ptest25+ Ptest75+ Ptest25- Ptest75-
section in. in. in. in. holes kips kips kips kips kips kips kips kips kips kips kips1 600S250-97 24 4.59 4.59 12 2 68.5 41.1 0.60 52.3 54.4 179.1 30.4 30.8 NaN NaN NaN NaN D2 600S162-97 24 3.79 3.79 12 2 56.5 33.9 0.60 43.1 38.4 99.4 25.6 25.3 NaN NaN NaN NaN D3 800S250-97 24 5.39 5.39 12 2 80.4 48.2 0.60 34.7 38.2 227.6 36.7 37.1 NaN NaN NaN NaN D4 800S200-97 24 4.99 4.99 12 2 74.4 44.6 0.60 32.4 31.6 169.9 31.7 30 NaN NaN NaN NaN D5 800S162-97 24 4.59 4.59 12 2 68.5 41.1 0.60 27.6 22.9 111.7 28.3 27 NaN NaN NaN NaN D6 600S137-68 24 3.59 3.59 12 2 37.5 22.5 0.60 13.8 11.6 49.8 14.9 14.3 NaN NaN NaN NaN D7 1000S250-97 24 6.19 6.19 12 2 92.3 55.4 0.60 25.5 25.5 246.2 35.1 34.4 NaN NaN NaN NaN D8 1000S200-97 24 5.79 5.79 12 2 86.3 51.8 0.60 23.5 20.2 176.8 33.7 33.8 NaN NaN NaN NaN D9 800S162-68 24 4.69 4.69 12 2 48.9 29.4 0.60 10.4 10.4 82.1 18.7 19.3 NaN NaN NaN NaN D10 1000S162-97 24 5.39 5.39 12 2 80.4 48.2 0.60 19.4 14.6 113.4 30.1 29.7 NaN NaN NaN NaN D11 600S137-43 24 3.66 3.66 12 2 24.2 14.5 0.60 3.7 4.6 33.0 7.7 7.82 NaN NaN NaN NaN D12 1200S250-97 24 6.99 6.99 12 2 104.2 62.5 0.60 19.7 17.2 242.2 36.7 33.8 NaN NaN NaN NaN D13 1000S200-68 24 5.89 5.89 12 2 61.5 36.9 0.60 8.5 9.3 128.5 19.4 19.5 NaN NaN NaN NaN D14 1000S250-54 24 6.33 6.33 12 2 52.5 31.5 0.60 4.5 7.6 142.5 15.3 15.2 NaN NaN NaN NaN D15 1200S162-97 24 6.19 6.19 12 2 92.3 55.4 0.60 14.7 9.9 108.3 31.2 32.4 NaN NaN NaN NaN D16 1000S162-68 24 5.49 5.49 12 2 57.3 34.4 0.60 7.6 6.5 83.6 17.9 17.7 NaN NaN NaN NaN D17 800S162-43 24 4.76 4.76 12 2 31.5 18.9 0.60 2.7 3.8 54.0 8.9 8.8 NaN NaN NaN NaN D18 1200S250-68 24 7.09 7.09 12 2 74.0 44.4 0.60 7.0 7.9 174.2 21.9 19.9 NaN NaN NaN NaN D19 1200S200-68 24 6.69 6.69 12 2 69.8 41.9 0.60 6.7 6.9 124.1 21.0 18.4 NaN NaN NaN NaN D20 1000S200-43 24 5.96 5.96 12 2 39.4 23.6 0.60 2.2 3.8 83.8 10.2 10.2 NaN NaN NaN NaN D21 600S250-97 24 3.45 3.45 12 2 68.5 47.9 0.70 52.3 56.7 221.4 38.0 36 NaN NaN NaN NaN D22 600S162-97 24 2.85 2.85 12 2 56.5 39.6 0.70 43.1 40.6 116.2 30.3 28.1 NaN NaN NaN NaN D23 800S250-97 24 4.05 4.05 12 2 80.4 56.3 0.70 34.7 40.0 294.5 43.0 42.4 NaN NaN NaN NaN D24 800S200-97 24 3.75 3.75 12 2 74.4 52.1 0.70 32.4 33.4 213.2 37.8 37.6 NaN NaN NaN NaN D25 800S162-97 24 3.45 3.45 12 2 68.5 47.9 0.70 27.6 24.6 136.4 32.4 30.7 NaN NaN NaN NaN D26 600S137-68 24 2.69 2.69 12 2 37.5 26.2 0.70 13.8 12.5 57.5 16.9 16.1 NaN NaN NaN NaN D27 1000S250-97 24 4.65 4.65 12 2 92.3 64.6 0.70 25.5 27.0 338.4 41.0 41.1 NaN NaN NaN NaN D28 1000S200-97 24 4.35 4.35 12 2 86.3 60.4 0.70 23.5 21.6 234.8 37.2 35.1 NaN NaN NaN NaN D29 800S162-68 24 3.51 3.51 12 2 48.9 34.3 0.70 10.4 11.0 100.9 19.3 18.5 NaN NaN NaN NaN D30 1000S162-97 24 4.05 4.05 12 2 80.4 56.3 0.70 19.4 16.0 146.0 34.4 38.4 NaN NaN NaN NaN D31 600S137-43 24 2.75 2.75 12 2 24.2 16.9 0.70 3.7 4.9 38.3 9.1 9.5 NaN NaN NaN NaN D32 1200S250-97 24 5.25 5.25 12 2 104.2 72.9 0.70 19.7 18.4 358.9 38.4 34 NaN NaN NaN NaN D33 1000S200-68 24 4.41 4.41 12 2 61.5 43.0 0.70 8.5 9.8 172.0 22.0 21.6 NaN NaN NaN NaN D34 1000S250-54 24 4.75 4.75 12 2 52.5 36.7 0.70 4.5 7.9 198.4 16.3 15.4 NaN NaN NaN NaN D35 1200S162-97 24 4.65 4.65 12 2 92.3 64.6 0.70 14.7 11.1 148.6 33.9 34.8 NaN NaN NaN NaN D36 1000S162-68 24 4.11 4.11 12 2 57.3 40.1 0.70 7.6 7.0 108.4 18.6 18.8 NaN NaN NaN NaN D37 800S162-43 24 3.57 3.57 12 2 31.5 22.0 0.70 2.7 4.0 66.8 9.8 9.75 NaN NaN NaN NaN D38 1200S250-68 24 5.31 5.31 12 2 74.0 51.8 0.70 7.0 8.4 260.8 23.4 23.3 NaN NaN NaN NaN D39 1200S200-68 24 5.01 5.01 12 2 69.8 48.9 0.70 6.7 7.3 178.2 21.6 20.1 NaN NaN NaN NaN D40 600S250-97 24 2.30 2.30 12 2 68.5 54.8 0.80 52.3 58.9 261.9 43.8 41.4 NaN NaN NaN NaN D41 600S162-97 24 1.90 1.90 12 2 56.5 45.2 0.80 43.1 42.8 132.5 34.3 30.3 NaN NaN NaN NaN D42 800S250-97 24 2.70 2.70 12 2 80.4 64.3 0.80 34.7 41.8 362.1 46.0 39.6 NaN NaN NaN NaN D43 800S200-97 24 2.50 2.50 12 2 74.4 59.5 0.80 32.4 35.1 256.9 41.7 40.1 NaN NaN NaN NaN D44 800S162-97 24 2.30 2.30 12 2 68.5 54.8 0.80 27.6 26.2 151.2 34.2 31.6 NaN NaN NaN NaN D45 600S137-68 24 1.79 1.79 12 2 37.5 30.0 0.80 13.8 13.3 62.4 18.1 16.4 NaN NaN NaN NaN D46 1000S250-97 24 3.10 3.10 12 2 92.3 73.8 0.80 25.5 28.4 434.3 44.9 41 NaN NaN NaN NaN D47 1000S200-97 24 2.90 2.90 12 2 86.3 69.1 0.80 23.5 23.0 294.8 38.9 36.4 NaN NaN NaN NaN D48 800S162-68 24 2.34 2.34 12 2 48.9 39.2 0.80 10.4 11.6 116.4 20.3 19.2 NaN NaN NaN NaN D49 1000S162-97 24 2.70 2.70 12 2 80.4 64.3 0.80 19.4 17.3 159.0 35.2 40.1 NaN NaN NaN NaN D50 600S137-43 24 1.83 1.83 12 2 24.2 19.4 0.80 3.7 5.1 43.3 9.5 9.96 NaN NaN NaN NaN D51 1200S250-97 24 3.50 3.50 12 2 104.2 83.4 0.80 19.7 19.6 483.4 39.7 36.3 NaN NaN NaN NaN D52 1000S200-68 24 2.94 2.94 12 2 61.5 49.2 0.80 8.5 10.3 217.2 22.9 22.3 NaN NaN NaN NaN D53 1000S250-54 24 3.17 3.17 12 2 52.5 42.0 0.80 4.5 8.2 256.8 16.0 15.3 NaN NaN NaN NaN D54 1200S162-97 24 3.10 3.10 12 2 92.3 73.8 0.80 14.7 12.4 164.7 34.5 35.2 NaN NaN NaN NaN D55 1000S162-68 24 2.74 2.74 12 2 57.3 45.8 0.80 7.6 7.4 122.4 19.7 18.7 NaN NaN NaN NaN D56 800S162-43 24 2.38 2.38 12 2 31.5 25.2 0.80 2.7 4.1 79.3 10.0 10 NaN NaN NaN NaN D57 1200S200-68 24 3.34 3.34 12 2 69.8 55.9 0.80 6.7 7.6 235.6 22.1 22.6 NaN NaN NaN NaN D58 1000S200-43 24 2.98 2.98 12 2 39.4 31.5 0.80 2.2 4.1 143.1 10.5 11 NaN NaN NaN NaN D59 600S250-97 24 1.15 1.15 12 2 68.5 61.6 0.90 52.3 61.2 299.2 48.1 42.6 NaN NaN NaN NaN D60 600S162-97 24 0.95 0.95 12 2 56.5 50.9 0.90 43.1 45.0 142.2 36.4 31 NaN NaN NaN NaN D61 800S250-97 24 1.35 1.35 12 2 80.4 72.3 0.90 34.7 43.6 426.6 47.6 43.7 NaN NaN NaN NaN D62 800S200-97 24 1.25 1.25 12 2 74.4 67.0 0.90 32.4 36.8 289.4 38.8 40.2 NaN NaN NaN NaN D63 800S162-97 24 1.15 1.15 12 2 68.5 61.6 0.90 27.6 27.8 153.2 34.8 32 NaN NaN NaN NaN D64 600S137-68 24 0.90 0.90 12 2 37.5 33.7 0.90 13.8 14.1 63.1 18.2 16.6 NaN NaN NaN NaN D65 1000S250-97 24 1.55 1.55 12 2 92.3 83.1 0.90 25.5 29.8 527.9 44.8 41 NaN NaN NaN NaN D66 1000S200-97 24 1.45 1.45 12 2 86.3 77.7 0.90 23.5 24.3 306.2 39.3 36.7 NaN NaN NaN NaN D67 800S162-68 24 1.17 1.17 12 2 48.9 44.0 0.90 10.4 12.2 118.0 20.5 19.5 NaN NaN NaN NaN D68 1000S162-97 24 1.35 1.35 12 2 80.4 72.3 0.90 19.4 18.6 160.9 35.7 40.6 NaN NaN NaN NaN D69 600S137-43 24 0.92 0.92 12 2 24.2 21.8 0.90 3.7 5.3 43.8 9.9 9.97 NaN NaN NaN NaN D70 1200S250-97 24 1.75 1.75 12 2 104.2 93.8 0.90 19.7 20.8 563.0 39.8 36.6 NaN NaN NaN NaN D71 1000S200-68 24 1.47 1.47 12 2 61.5 55.3 0.90 8.5 10.8 231.0 22.6 22.5 NaN NaN NaN NaN D72 1000S250-54 24 1.58 1.58 12 2 52.5 47.2 0.90 4.5 8.5 313.9 16.1 15.4 NaN NaN NaN NaN D73 1200S162-97 24 1.55 1.55 12 2 92.3 83.1 0.90 14.7 13.6 166.7 34.7 35.3 NaN NaN NaN NaN D74 1000S162-68 24 1.37 1.37 12 2 57.3 51.6 0.90 7.6 7.9 124.0 19.9 18.9 NaN NaN NaN NaN D75 800S162-43 24 1.19 1.19 12 2 31.5 28.3 0.90 2.7 4.3 80.4 10.1 10.1 NaN NaN NaN NaN D76 1200S250-68 24 1.77 1.77 12 2 74.0 66.6 0.90 7.0 9.2 419.7 22.5 23.6 NaN NaN NaN NaN D77 1200S200-68 24 1.67 1.67 12 2 69.8 62.9 0.90 6.7 8.0 240.7 24.2 22.8 NaN NaN NaN NaN D78 1000S200-43 24 1.49 1.49 12 2 39.4 35.5 0.90 2.2 4.3 155.1 10.5 11.1 NaN NaN NaN NaN D79 600S250-97 24 4.59 0.00 12 2 68.5 68.5 1.00 52.3 63.0 332.0 50.4 46.8 NaN NaN NaN NaN D80 600S162-97 24 3.79 0.00 12 2 56.5 56.5 1.00 43.1 47.1 142.7 38.7 34 NaN NaN NaN NaN D81 800S250-97 24 5.39 0.00 12 2 80.4 80.4 1.00 34.7 45.4 484.1 48.7 45.9 NaN NaN NaN NaN D82 800S200-97 24 4.99 0.00 12 2 74.4 74.4 1.00 32.4 38.4 290.8 40.8 37 NaN NaN NaN NaN D83 800S162-97 24 4.59 0.00 12 2 68.5 68.5 1.00 27.6 29.4 153.7 35.1 32.1 NaN NaN NaN NaN D84 600S137-68 24 3.59 0.00 12 2 37.5 37.5 1.00 13.8 14.9 63.3 18.3 16.6 NaN NaN NaN NaN D85 1000S250-97 24 6.19 0.00 12 2 92.3 92.3 1.00 25.5 31.2 541.3 44.5 43.1 NaN NaN NaN NaN D86 1000S200-97 24 5.79 0.00 12 2 86.3 86.3 1.00 23.5 25.6 307.6 44.3 42.6 NaN NaN NaN NaN D87 800S162-68 24 4.69 0.00 12 2 48.9 48.9 1.00 10.4 12.8 118.4 22.9 24.9 NaN NaN NaN NaN D88 1000S162-97 24 5.39 0.00 12 2 80.4 80.4 1.00 19.4 19.9 161.5 35.4 33.7 NaN NaN NaN NaN D89 600S137-43 24 3.66 0.00 12 2 24.2 24.2 1.00 3.7 5.5 44.0 9.6 9.27 NaN NaN NaN NaN D90 1200S250-97 24 6.99 0.00 12 2 104.2 104.2 1.00 19.7 22.0 566.1 45.7 40.3 NaN NaN NaN NaN D91 1000S200-68 24 5.89 0.00 12 2 61.5 61.5 1.00 8.5 11.3 232.1 21.9 20.5 NaN NaN NaN NaN D92 1000S250-54 24 6.33 0.00 12 2 52.5 52.5 1.00 4.5 8.8 329.7 16.6 15.7 NaN NaN NaN NaN D93 1200S162-97 24 6.19 0.00 12 2 92.3 92.3 1.00 14.7 14.8 167.2 34.7 35.3 NaN NaN NaN NaN D94 1000S162-68 24 5.49 0.00 12 2 57.3 57.3 1.00 7.6 8.3 124.4 20.0 19.8 NaN NaN NaN NaN D95 800S162-43 24 4.76 0.00 12 2 31.5 31.5 1.00 2.7 4.4 80.7 10.3 10 NaN NaN NaN NaN D96 1200S250-68 24 7.09 0.00 12 2 74.0 74.0 1.00 7.0 9.6 422.1 21.9 24.7 NaN NaN NaN NaN D97 1200S200-68 24 6.69 0.00 12 2 69.8 69.8 1.00 6.7 8.4 241.8 24.2 22.8 NaN NaN NaN NaN D98 1000S200-43 24 5.96 0.00 12 2 39.4 39.4 1.00 2.2 4.4 155.9 11.1 10.9 NaN NaN NaN NaN D99 1200S200-97 26 3.30 3.30 13 2 98.2 78.6 0.80 17.8 17.3 267.6 35.7 31.3 NaN NaN NaN NaN
Study Type
526
ID # SSMA L Lhole hhole S # of Pyg Pynet Pynet/Pyg Pcrl Pcrd Pcre Ptest25 Ptest75 Ptest25+ Ptest75+ Ptest25- Ptest75-
section in. in. in. in. holes kips kips kips kips kips kips kips kips kips kips kips100 1200S162-54 24 3.17 3.17 12 2 52.5 42.0 0.80 3.1 3.0 105.2 13.5 13.5 NaN NaN NaN NaN101 1200S250-68 46 3.54 3.54 15 3 74.0 59.2 0.80 7.0 8.8 96.9 22.6 22.4 NaN NaN NaN NaN102 1000S200-68 42 2.94 2.94 14 3 61.5 49.2 0.80 8.5 10.3 71.7 21.3 21 NaN NaN NaN NaN103 1000S162-54 36 2.77 2.77 12 3 45.9 36.7 0.80 4.0 4.5 45.1 12.4 12.5 NaN NaN NaN NaN104 800S137-68 32 2.19 2.19 16 2 45.8 36.7 0.80 8.8 7.9 37.5 16.9 17.2 NaN NaN NaN NaN105 1000S162-43 44 2.78 2.78 14 3 36.7 29.4 0.80 2.0 2.5 24.8 8.3 8.68 NaN NaN NaN NaN106 1200S162-54 46 3.17 3.17 15 3 52.5 42.0 0.80 3.1 3.0 28.6 10.0 10.3 NaN NaN NaN NaN107 250S137-54 26 4.00 1.11 13 2 18.5 14.8 0.80 24.9 19.8 12.8 11.2 10.2 10.6 9.51 NaN NaN108 250S137-54 32 4.00 1.11 16 2 18.5 14.8 0.80 24.9 19.8 9.5 10.9 10.1 9.71 8.51 NaN NaN G109 400S162-68 54 4.00 1.54 13 4 32.2 25.8 0.80 27.4 30.4 11.6 13.7 12.1 13.2 11.4 NaN NaN G110 600S250-97 92 4.00 2.30 13 7 68.5 54.8 0.80 52.3 55.6 18.8 24.4 22.2 23.2 20.5 NaN NaN G111 350S162-54 66 4.00 1.47 13 5 24.3 19.4 0.80 16.7 20.6 5.3 7.9 7.14 6.8 6.27 NaN NaN G112 250S162-33 58 4.00 1.29 14 4 13.1 10.5 0.80 6.3 8.8 2.4 3.1 4.29 2.86 2.59 NaN NaN G113 250S137-33 60 4.00 1.14 12 5 11.6 9.2 0.80 5.7 7.2 1.6 3.2 2.91 2.16 2.04 NaN NaN G114 362S137-43 84 4.00 1.35 12 7 17.9 14.3 0.80 7.4 9.7 2.1 2.6 2.37 2.63 2.37 2.63 2.37 G115 362S137-68 88 4.00 1.32 12 7 27.5 22.0 0.80 28.8 24.0 3.9 3.7 3.37 3.75 3.2 3.63 3.23 G116 250S162-54 96 4.00 1.27 12 8 21.0 16.8 0.80 27.4 23.5 2.0 3.5 3.29 2.54 2.46 NaN NaN G117 600S137-54 96 4.00 1.81 12 8 30.1 24.1 0.80 7.2 7.1 3.2 2.9 2.64 3.03 2.91 2.76 2.53 G118 250S137-33 94 4.00 1.14 13 7 11.6 9.2 0.80 5.7 7.2 0.9 1.5 1.36 1.07 1.03 NaN NaN G119 800S137-97 94 4.00 2.15 13 7 64.0 51.2 0.80 22.9 15.8 5.5 5.0 4.6 5.09 4.83 4.83 4.41 G120 800S137-97 96 4.00 2.15 12 8 64.0 51.2 0.80 22.9 15.8 5.2 4.7 4.4 4.85 4.6 4.61 4.21 G121 250S137-68 12 1.09 1.09 12 1 22.8 18.3 0.80 49.5 38.5 89.9 16.9 15.8 NaN NaN NaN NaN G122 250S162-68 16 1.24 1.24 16 1 26.0 20.8 0.80 54.7 43.6 68.9 19.3 19.3 NaN NaN NaN NaN G123 250S162-68 22 1.24 1.24 22 1 26.0 20.8 0.80 54.7 43.6 37.4 18.2 17.8 NaN NaN NaN NaN G124 250S137-54 26 4.00 0.56 13 2 18.5 16.6 0.90 24.9 19.8 13.2 12.8 11.2 12.30 10.20 NaN NaN G125 250S137-54 32 4.00 0.56 16 2 18.5 16.6 0.90 24.9 19.8 9.8 11.7 10.2 10.20 8.66 NaN NaN G126 400S162-68 54 4.00 0.77 13 4 32.2 29.0 0.90 27.4 30.4 12.0 15.2 12.6 14.40 12.00 NaN NaN G127 600S250-97 92 4.00 1.15 13 7 68.5 61.6 0.90 52.3 55.6 19.5 25.6 22.8 23.20 20.50 NaN NaN G128 250S162-33 58 4.00 0.64 14 4 13.1 11.8 0.90 6.3 8.8 2.4 3.2 3.09 2.91 2.62 NaN NaN G129 250S137-33 60 4.00 0.57 12 5 11.6 10.4 0.90 5.7 7.2 1.7 2.4 2.63 2.22 2.07 NaN NaN G130 362S137-43 84 4.00 0.68 12 7 17.9 16.1 0.90 7.4 9.7 2.2 2.8 2.4 2.76 2.41 2.76 2.41 G131 362S137-68 88 4.00 0.66 12 7 27.5 24.8 0.90 28.5 24.0 4.0 3.9 3.47 3.85 3.26 3.83 3.32 G132 250S162-54 96 4.00 0.63 12 8 21.0 18.9 0.90 27.4 23.5 2.1 4.0 3.51 2.61 2.52 NaN NaN G133 600S137-54 96 4.00 0.91 12 8 30.1 27.1 0.90 7.0 7.1 3.3 3.0 2.67 3.22 3.07 2.83 2.55 G134 250S137-33 94 4.00 0.57 13 7 11.6 10.4 0.90 5.7 7.2 0.9 1.6 1.39 1.10 1.06 NaN NaN G135 800S137-97 94 4.00 1.07 13 7 64.0 57.6 0.90 22.9 15.8 5.5 5.0 4.64 5.18 4.87 4.90 4.44 G136 800S137-97 96 4.00 1.07 12 8 64.0 57.6 0.90 22.9 15.8 5.3 4.8 4.43 4.94 4.64 4.69 4.25 G137 250S162-68 8 0.62 0.62 8 1 26.0 23.4 0.90 54.7 44.7 280.3 24.0 21.3 NaN NaN NaN NaN G138 250S137-68 12 0.55 0.55 12 1 22.8 20.5 0.90 49.5 39.8 95.2 19.9 17 NaN NaN NaN NaN G139 250S162-68 16 0.62 0.62 16 1 26.0 23.4 0.90 54.7 44.7 74.5 23.5 21.2 NaN NaN NaN NaN G140 250S162-68 22 0.62 0.62 22 1 26.0 23.4 0.90 54.7 44.7 40.8 22.9 19.8 NaN NaN NaN NaN G141 250S162-68 8 4.00 0.00 8 1 26.0 26.0 1.00 54.7 45.0 303.2 24.7 21.5 24.7 21.5 NaN NaN G142 250S137-68 12 4.00 0.00 12 1 22.8 22.8 1.00 49.5 40.9 99.7 20.6 17.1 20.6 17.1 NaN NaN G143 250S162-68 16 4.00 0.00 16 1 26.0 26.0 1.00 54.7 45.0 77.3 23.0 21.4 23 21.4 NaN NaN G144 250S162-68 22 4.00 0.00 22 1 26.0 26.0 1.00 54.7 45.0 41.9 22.8 19.9 22.8 19.9 NaN NaN G145 250S137-54 26 4.00 0.00 13 2 18.5 18.5 1.00 24.9 24.7 18.5 15.1 11.7 13.4 10.7 NaN NaN G146 250S137-54 32 4.00 0.00 16 2 18.5 18.5 1.00 24.9 24.7 12.7 11.9 10.5 10.5 8.94 NaN NaN G147 400S162-68 54 4.00 0.00 13 4 32.2 32.2 1.00 27.4 35.6 16.4 16.3 13.4 14.6 12.8 NaN NaN G148 600S250-97 92 4.00 0.00 13 7 68.5 68.5 1.00 52.3 63.0 26.7 25.6 23.3 23.4 20.8 NaN NaN G149 350S162-54 66 4.00 0.00 13 5 24.3 24.3 1.00 16.7 24.0 7.5 7.5 7.65 7 6.52 NaN NaN G150 250S162-33 58 4.00 0.00 14 4 13.1 13.1 1.00 6.3 9.8 3.3 5.4 3.06 2.98 2.72 NaN NaN G151 250S137-33 60 4.00 0.00 12 5 11.6 11.6 1.00 5.7 8.7 2.4 2.4 2.41 2.27 2.15 NaN NaN G152 362S137-43 84 4.00 0.00 12 7 17.9 17.9 1.00 7.4 11.6 3.1 3.0 2.48 2.76 2.48 2.61 2.48 G153 362S137-68 88 4.00 0.00 12 7 27.5 27.5 1.00 28.8 31.5 4.1 4.0 3.59 3.9 3.31 3.99 3.43 G154 250S162-54 96 4.00 0.00 12 8 21.0 21.0 1.00 27.4 27.8 2.7 4.2 3.63 2.68 2.61 NaN NaN G155 600S137-54 96 4.00 0.00 12 8 30.1 30.1 1.00 7.2 8.9 3.3 3.1 2.76 3.25 3.09 2.94 2.62 G156 250S137-33 94 4.00 0.00 13 7 11.6 11.6 1.00 5.7 8.7 1.1 1.2 1.42 1.12 1.1 NaN NaN G157 800S137-97 94 4.00 0.00 13 7 64.0 64.0 1.00 22.9 22.9 5.6 5.1 4.71 5.32 4.94 5.01 4.5 G158 800S137-97 96 4.00 0.00 12 8 64.0 64.0 1.00 22.9 22.9 5.3 4.9 4.5 5.08 4.71 4.81 4.3 G159 350S162-68 34 2.53 2.53 17 2 30.1 19.6 0.65 33.5 35.2 25.9 13.1 12.9 13 12.7 13 12.7 L160 1000S200-97 88 5.07 5.07 12 7 86.3 56.1 0.65 23.5 20.9 18.6 15.7 15.3 16.2 16.2 15.2 14.5 L161 350S162-54 24 2.56 2.56 12 2 24.3 15.8 0.65 16.7 21.9 36.6 11.2 11 11.2 11 11.2 11 L162 800S200-68 74 4.45 4.45 12 6 53.1 34.5 0.65 11.4 15.2 16.0 12.6 12.4 12.2 11.8 12.9 13.1 L163 550S162-54 42 3.26 3.26 14 3 30.9 20.1 0.65 8.8 13.0 21.5 10.7 10.2 10.9 10.6 10.4 9.86 L165 800S200-54 66 4.49 4.49 13 5 42.5 27.6 0.65 5.8 9.6 16.4 10.2 10.2 9.97 9.85 10.4 10.6 L166 600S250-43 56 4.17 4.17 14 4 31.5 20.5 0.65 4.6 10.6 19.1 9.6 9.81 9.51 9.01 9.77 9.93 L167 600S162-43 32 3.47 3.47 16 2 26.2 17.0 0.65 4.0 6.8 33.0 8.8 8.8 8.83 9.04 8.77 8.72 L168 800S250-43 74 4.87 4.87 12 6 36.7 23.9 0.65 3.1 6.9 13.4 8.7 8.68 8.87 8.9 8.58 8.39 L169 800S162-43 40 4.17 4.17 13 3 31.5 20.5 0.65 2.7 3.9 23.4 8.3 8.33 7.89 8.18 NaN 6.88 L170 1000S250-43 80 5.57 5.57 13 6 42.0 27.3 0.65 2.3 4.5 13.8 8.2 7.74 NaN 8 8.11 7.46 L171 350S162-68 34 1.44 1.44 17 2 30.1 24.1 0.80 33.5 37.0 28.7 19.7 18 19.3 17.4 18.3 17.4 L172 1000S200-97 88 2.90 2.90 12 7 86.3 69.1 0.80 23.5 23.0 22.4 17.0 16.1 17.9 17.3 16.3 15.2 L173 350S162-54 24 1.47 1.47 12 2 24.3 19.4 0.80 16.7 22.9 42.9 15.2 14.3 15.1 14.3 15.1 14.3 L174 800S200-68 74 2.54 2.54 12 6 53.1 42.5 0.80 11.4 16.1 21.3 13.9 13.6 13.5 13 14.3 14.3 L175 550S162-54 42 1.87 1.87 14 3 30.9 24.7 0.80 8.8 13.7 25.4 12.7 11.9 13 12.3 12.5 11.5 L176 800S200-54 66 2.57 2.57 13 5 42.5 34.0 0.80 5.8 10.1 21.5 11.2 11.2 11 10.7 11.5 11.6 L177 600S250-43 56 2.38 2.38 14 4 31.5 25.2 0.80 4.6 11.0 24.1 11.7 11.6 11.6 11.4 11.9 11.9 L178 600S162-43 32 1.98 1.98 16 2 26.2 20.9 0.80 4.0 7.2 38.6 10.1 10 10.2 10.1 9.98 9.89 L179 800S250-43 74 2.78 2.78 12 6 36.7 29.4 0.80 3.1 7.3 18.6 8.8 9.43 9.38 9.71 NaN 9.19 L180 800S162-43 40 2.38 2.38 13 3 31.5 25.2 0.80 2.7 4.1 28.6 8.1 8.82 8.46 8.68 9.14 9.25 L181 1000S250-43 80 3.18 3.18 13 6 42.0 33.6 0.80 2.3 4.8 19.7 8.8 8.78 8.14 8.84 8.69 8.62 L182 400S162-68 42 4.00 1.54 14 3 32.2 25.8 0.80 27.4 30.4 16.9 17.5 15.7 NaN NaN NaN NaN183 250S137-33 32 4.00 1.14 16 2 11.6 9.2 0.80 5.7 7.2 5.0 5.6 5.33 NaN NaN NaN NaN184 250S137-33 18 4.00 1.14 18 1 11.6 9.2 0.80 5.7 7.2 15.2 5.8 5.41 NaN NaN NaN NaN185 362S200-43 20 4.00 1.71 20 1 22.5 18.0 0.80 8.9 14.7 53.7 11.1 11.2 NaN NaN NaN NaN186 362S137-33 30 4.00 1.36 15 2 13.8 11.1 0.80 3.3 5.8 9.7 5.6 5.6 NaN NaN NaN NaN187 800S137-54 42 4.00 2.21 14 3 36.7 29.4 0.80 4.8 4.2 17.0 8.8 8.79 NaN NaN NaN NaN188 400S162-33 18 4.00 1.59 18 1 16.1 12.9 0.80 3.2 6.9 42.7 6.8 7.06 NaN NaN NaN NaN189 600S162-43 26 4.00 1.98 13 2 26.2 20.9 0.80 4.0 6.7 42.8 10.7 10.7 NaN NaN NaN NaN190 800S250-54 34 4.00 2.77 17 2 45.9 36.7 0.80 6.1 11.9 90.9 17.9 16.7 NaN NaN NaN NaN191 250S137-68 10 4.00 1.09 10 1 22.8 18.3 0.80 49.5 31.1 92.2 14.8 13.2 NaN NaN NaN NaN192 250S137-68 12 4.00 1.09 12 1 22.8 18.3 0.80 49.5 31.1 64.8 15.1 13.8 NaN NaN NaN NaN193 250S162-68 16 4.00 1.24 16 1 26.0 20.8 0.80 54.7 38.2 50.7 17.6 17.3 NaN NaN NaN NaN194 250S137-54 18 4.00 1.11 18 1 18.5 14.8 0.80 24.9 19.8 24.3 11.9 11.1 NaN NaN NaN NaN195 250S162-68 22 4.00 1.24 22 1 26.0 20.8 0.80 54.7 38.2 27.7 16.7 16.5 NaN NaN NaN NaN196 250S137-68 24 4.00 1.09 12 2 22.8 18.3 0.80 49.5 31.1 18.0 14.3 13.8 NaN NaN NaN NaN197 362S200-68 40 4.00 1.67 13 3 34.8 27.9 0.80 35.3 37.8 21.6 20.2 18.7 NaN NaN NaN NaN198 350S162-68 40 4.00 1.44 13 3 30.1 24.1 0.80 33.5 32.7 15.6 16.4 14.7 NaN NaN NaN NaN199 250S137-68 34 4.00 1.09 17 2 22.8 18.3 0.80 49.5 31.1 10.2 12.7 11.9 NaN NaN NaN NaN
Study Type
527
ID # SSMA L Lhole hhole S # of Pyg Pynet Pynet/Pyg Pcrl Pcrd Pcre Ptest25 Ptest75 Ptest25+ Ptest75+ Ptest25- Ptest75-
section in. in. in. in. holes kips kips kips kips kips kips kips kips kips kips kips200 362S137-68 46 4.00 1.32 15 3 27.5 22.0 0.80 28.8 24.0 9.9 11.9 10.8 NaN NaN NaN NaN201 250S162-43 42 4.00 1.28 14 3 16.9 13.5 0.80 13.9 14.9 5.3 7.2 6.67 NaN NaN NaN NaN202 350S162-68 34 4.00 0.00 17 2 30.1 30.1 1.00 33.5 39.4 31.5 22.6 19.4 22.3 18.7 19.3 18.7 L203 1000S200-97 88 4.00 0.00 12 7 86.3 86.3 1.00 23.5 25.6 22.9 17.8 16.5 19.4 18 17 15.5 L204 350S162-54 24 4.00 0.00 12 2 24.3 24.3 1.00 16.7 24.0 49.4 16.9 15 16.9 15 16.9 15 L205 800S200-68 74 4.00 0.00 12 6 53.1 53.1 1.00 11.4 17.4 23.1 14.1 13.9 14 13.2 14.7 14.5 L206 550S162-54 42 4.00 0.00 14 3 30.9 30.9 1.00 8.8 14.6 29.0 13.0 12 13.2 12.5 12.7 11.6 L208 800S200-54 66 4.00 0.00 13 5 42.5 42.5 1.00 5.8 10.8 23.8 11.6 11.4 11.3 11 11.8 11.9 L209 600S250-43 56 4.00 0.00 14 4 31.5 31.5 1.00 4.6 11.1 29.6 12.2 12 12 11.8 12.3 12.2 L210 600S162-43 32 4.00 0.00 16 2 26.2 26.2 1.00 4.0 7.6 42.1 10.1 10.1 10.2 10.3 10 9.97 L211 800S250-43 74 4.00 0.00 12 6 36.7 36.7 1.00 3.1 7.7 24.9 9.7 9.64 9.84 9.9 9.62 9.4 L212 800S162-43 40 4.00 0.00 13 3 31.5 31.5 1.00 2.7 4.4 29.1 8.7 8.94 NaN 8.74 9.32 9.42 L213 1000S250-43 80 4.00 0.00 13 6 42.0 42.0 1.00 2.3 5.1 24.2 8.9 8.89 8.82 9.64 8.79 8.72 L
Study Type
528
Appendix L 20BSimulated beam experiment database
The table provided in this appendix summarizes the dimensions, elastic buckling
moments, and tested strengths of simulated beam experiments described in Section
1858H8.2.1. The letters in the “Study type” column denote simulations considered in the DSM
failure mode studies: “D” for the distortional buckling failure study in Section 1859H8.2.3 and
“L” for the local buckling failure study in Section 1860H8.2.2.
529
ID # SSMA section L Lhole hhole # of holes S Fy My Mynet Mcrl Mcrd Mtest25 Mtest75 Study Typein. in. in. in. ksi kip·in. kip·in. kip·in. kip·in. kip·in. kip·in.
1 400S162-68 48 3.1 3.1 3 16 58.6 39.4 34.3 156.3 77.1 35.1 30.22 400S137-54 48 2.8 2.8 3 16 58.6 27.9 24.8 76.5 40.8 24.4 22.63 550S162-54 48 3.7 3.7 3 16 58.6 49.5 44.3 74.8 57.1 41.1 39.24 800S137-68 48 4.4 4.4 3 16 58.6 92.2 84.9 101.6 68.8 73.2 66.25 800S162-54 48 4.7 4.7 3 16 58.6 84.0 76.7 62.7 61.5 62.3 61.46 800S137-54 48 4.4 4.4 3 16 58.6 74.8 68.8 52.2 43.0 55.9 50.97 1200S250-68 48 7.1 7.1 3 16 58.6 239.0 218.3 124.9 123.7 155 1528 1200S162-68 48 6.3 6.3 3 16 58.6 190.5 176.1 87.6 72.1 134 1289 1200S250-54 48 7.1 7.1 3 16 58.6 192.1 175.4 63.3 73.9 108 107
10 1000S162-43 48 5.6 5.6 3 16 58.6 94.0 86.4 26.6 32.7 58.1 56.411 1200S162-54 48 6.3 6.3 3 16 58.6 153.5 141.9 44.6 45.7 98.7 96.212 400S162-68 48 2.3 2.3 3 16 58.6 39.4 37.2 156.3 79.8 36.7 31.213 400S137-54 48 2.1 2.1 3 16 58.6 27.9 26.6 69.6 42.5 26.3 23.314 550S162-54 48 2.8 2.8 3 16 58.6 49.5 47.3 62.7 59.3 44.4 40.715 800S137-68 48 3.3 3.3 3 16 58.6 92.2 89.1 83.0 75.7 74.8 67.316 800S162-54 48 3.5 3.5 3 16 58.6 84.0 80.9 47.1 64.7 67.4 63.817 800S137-54 48 3.3 3.3 3 16 58.6 74.8 72.3 42.1 46.4 57.1 52.118 1200S250-68 48 5.3 5.3 3 16 58.6 239.0 230.3 93.3 130.3 159 15519 1200S162-68 48 4.7 4.7 3 16 58.6 190.5 184.4 76.9 79.4 137 12920 1200S250-54 48 5.3 5.3 3 16 58.6 192.1 185.1 47.1 77.4 105 10921 1000S162-43 48 4.2 4.2 3 16 58.6 94.0 90.8 21.3 34.6 53.8 57.722 1200S162-54 48 4.7 4.7 3 16 58.6 153.5 148.6 38.7 49.2 99.9 99.123 400S162-68 48 1.5 1.5 3 16 58.6 39.4 38.8 148.4 82.3 39.3 31.324 400S137-54 48 1.4 1.4 3 16 58.6 27.9 27.5 71.8 44.2 26.5 23.425 550S162-54 48 1.9 1.9 3 16 58.6 49.5 48.8 68.6 61.5 44.6 41.126 800S137-68 48 2.2 2.2 3 16 58.6 92.2 91.3 105.7 82.3 75.6 67.827 800S162-54 48 2.4 2.4 3 16 58.6 84.0 83.1 62.7 67.8 68.3 64.528 800S137-54 48 2.2 2.2 3 16 58.6 74.8 74.0 55.4 49.6 57.5 52.929 1200S250-68 48 3.5 3.5 3 16 58.6 239.0 236.4 124.9 136.7 162 15730 1200S162-68 48 3.1 3.1 3 16 58.6 190.5 188.7 99.6 86.5 136 13031 1200S250-54 48 3.6 3.6 3 16 58.6 192.1 190.0 63.3 80.8 114 11232 1000S162-43 48 2.8 2.8 3 16 58.6 94.0 93.1 29.0 36.4 56 58.533 1200S162-54 48 3.2 3.2 3 16 58.6 153.5 152.1 51.7 52.7 100 99.734 400S162-68 48 0.8 0.8 3 16 58.6 39.4 39.3 156.3 84.7 39.5 31.335 400S137-54 48 0.7 0.7 3 16 58.6 27.9 27.8 76.5 45.8 26.5 23.436 550S162-54 48 0.9 0.9 3 16 58.6 49.5 49.4 74.8 63.6 44.9 41.337 800S137-68 48 1.1 1.1 3 16 58.6 92.2 92.1 105.7 88.7 76 6838 800S162-54 48 1.2 1.2 3 16 58.6 84.0 83.9 62.7 70.9 68.7 65.339 800S137-54 48 1.1 1.1 3 16 58.6 74.8 74.7 55.4 52.8 57.6 53.240 1200S250-68 48 1.8 1.8 3 16 58.6 239.0 238.7 124.9 143.1 163 16041 1200S162-68 48 1.6 1.6 3 16 58.6 190.5 190.3 99.6 93.3 136 13042 1200S250-54 48 1.8 1.8 3 16 58.6 192.1 191.9 63.3 84.2 111 11243 1000S162-43 48 1.4 1.4 3 16 58.6 94.0 93.9 29.0 38.1 56.9 58.844 1200S162-54 48 1.6 1.6 3 16 58.6 153.5 153.4 51.7 56.0 100 99.945 400S162-68 48 0.0 0.0 0 16 58.6 39.4 39.4 156.3 85.4 39.4 33.1 D46 400S137-54 48 0.0 0.0 0 16 58.6 27.9 27.9 76.5 46.6 25.4 22.6 D47 550S162-54 48 0.0 0.0 0 16 58.6 49.5 49.5 74.8 64.9 44.7 41.4 D48 800S137-68 48 0.0 0.0 0 16 58.6 92.2 92.2 105.7 94.2 76.2 68.2 D49 800S162-54 48 0.0 0.0 0 16 58.6 84.0 84.0 62.7 73.3 65.4 60.5 D50 800S137-54 48 0.0 0.0 0 16 58.6 74.8 74.8 55.4 55.6 57 52.8 D51 1200S250-68 48 0.0 0.0 0 16 58.6 239.0 239.0 124.9 148.4 NaN 159 D52 1200S162-68 48 0.0 0.0 0 16 58.6 190.5 190.5 99.6 99.4 136 130 D53 1200S250-54 48 0.0 0.0 0 16 58.6 192.1 192.1 63.3 87.0 NaN 119 D54 1000S162-43 48 0.0 0.0 0 16 58.6 94.0 94.0 29.0 39.7 59.8 61.9 D55 1200S162-54 48 0.0 0.0 0 16 58.6 153.5 153.5 51.7 59.1 100 100 D56 400S162-68 48 2.8 2.8 3 16 58.6 39.4 35.5 156.3 78.0 36.5 31.8 D57 400S137-54 48 2.7 2.7 3 16 58.6 27.9 25.1 76.5 41.1 25.3 23.3 D58 550S162-54 48 3.7 3.7 3 16 58.6 49.5 44.5 74.8 57.2 41.3 39.3 D59 800S137-68 48 4.7 4.7 3 16 58.6 92.2 83.0 105.7 66.5 71.7 65.6 D60 800S162-54 48 5.0 5.0 3 16 58.6 84.0 75.6 62.7 60.9 60.8 60.1 D61 800S137-54 48 4.8 4.8 3 16 58.6 74.8 67.3 55.4 42.0 55.2 50.6 D62 1200S250-68 48 7.4 7.4 3 16 58.6 239.0 215.1 124.9 122.4 158 168 D63 1200S162-68 48 6.9 6.9 3 16 58.6 190.5 171.4 99.6 69.1 133 127 D64 1200S250-54 48 7.5 7.5 3 16 58.6 192.1 172.9 63.3 73.3 108 107 D65 1000S162-43 48 6.0 6.0 3 16 58.6 94.0 84.6 29.0 32.2 57.4 51.7 D66 1200S162-54 48 6.9 6.9 3 16 58.6 153.5 138.2 50.9 44.3 97.6 95.3 D67 400S162-68 48 2.2 2.2 3 16 58.6 39.4 37.4 153.4 80.0 38.4 33.8 D68 400S137-54 48 2.2 2.2 3 16 58.6 27.9 26.5 71.1 42.5 26.2 23.3 D69 550S162-54 48 2.9 2.9 3 16 58.6 49.5 47.0 65.7 59.1 44.1 40.6 D70 800S137-68 48 3.8 3.8 3 16 58.6 92.2 87.6 85.0 72.8 74.3 66.9 D71 800S162-54 48 3.9 3.9 3 16 58.6 84.0 79.8 49.9 63.7 66.5 63.4 D72 800S137-54 48 3.8 3.8 3 16 58.6 74.8 71.1 43.3 45.0 56.8 52.1 D73 1200S250-68 48 5.9 5.9 3 16 58.6 239.0 227.0 99.0 128.1 158 154 D74 1200S162-68 48 5.5 5.5 3 16 58.6 190.5 181.0 76.9 75.9 136 129 D75 1200S250-54 48 5.9 5.9 3 16 58.6 192.1 182.5 50.1 76.3 104 109 D76 1000S162-43 48 4.7 4.7 3 16 58.6 94.0 89.3 22.1 33.8 55.2 56.8 D77 1200S162-54 48 5.5 5.5 3 16 58.6 153.5 145.9 38.9 47.6 99.4 96.5 D
530
ID # SSMA section L Lhole hhole # of holes S Fy My Mynet Mcrl Mcrd Mtest25 Mtest75 Study Typein. in. in. in. ksi kip·in. kip·in. kip·in. kip·in. kip·in. kip·in.
78 550S162-33 48 0.0 0.0 0 16 58.6 31.1 31.1 17.5 23.5 22.2 20.4 L79 600S162-33 48 0.0 0.0 0 16 58.6 35.0 35.0 16.8 24.0 22.9 23.6 L80 1000S200-54 48 0.0 0.0 0 16 58.6 132.1 132.1 63.2 85.9 94.4 88.8 L81 800S162-43 48 0.0 0.0 0 16 58.6 67.8 67.8 32.2 43.7 48.7 47.5 L82 800S200-43 48 0.0 0.0 0 16 58.6 77.6 77.6 35.9 53.1 54.2 51.4 L83 600S200-33 48 0.0 0.0 0 16 58.6 40.5 40.5 18.4 26.8 25.7 25.2 L84 1200S200-54 48 0.0 0.0 0 16 58.6 172.4 172.4 58.1 84.2 115 112 L85 1000S200-43 48 0.0 0.0 0 16 58.6 106.4 106.4 32.3 53.7 67.6 65.3 L86 1000S250-43 48 0.0 0.0 0 16 58.6 119.5 119.5 35.2 54.2 66.4 66.7 L87 800S137-33 48 0.0 0.0 0 16 58.6 46.8 46.8 13.3 19.2 26.3 26.7 L88 800S162-33 48 0.0 0.0 0 16 58.6 52.4 52.4 14.7 24.9 31.5 31.8 L89 800S200-33 48 0.0 0.0 0 16 58.6 60.0 60.0 16.3 31.0 33.8 32.3 L90 550S162-33 48 4.2 4.2 3 16 58.6 31.1 26.4 17.5 20.9 19 19 L91 600S162-33 48 4.5 4.5 3 16 58.6 35.0 29.7 16.8 21.2 20.2 20.3 L92 1000S200-54 48 7.1 7.1 3 16 58.6 132.1 112.3 63.2 70.5 81.9 82 L93 800S162-43 48 5.7 5.7 3 16 58.6 67.8 57.7 32.2 35.9 41.7 40.8 L94 800S200-43 48 6.0 6.0 3 16 58.6 77.6 66.0 35.9 46.2 45.6 45.5 L95 600S200-33 48 4.8 4.8 3 16 58.6 40.5 34.4 18.4 24.5 22.2 22.2 L96 1200S200-54 48 8.2 8.2 3 16 58.6 172.4 146.5 58.1 67.3 100 103 L97 1000S200-43 48 7.1 7.1 3 16 58.6 106.4 90.5 32.3 45.2 57.7 57.7 L98 1000S250-43 48 7.4 7.4 3 16 58.6 119.5 101.6 35.2 46.7 61 59.6 L99 800S137-33 48 5.5 5.5 3 16 58.6 46.8 39.8 13.3 15.3 25.5 25.6 L
100 800S162-33 48 5.7 5.7 3 16 58.6 52.4 44.6 14.7 21.1 27.6 27.5 L101 800S200-33 48 6.0 6.0 3 16 58.6 60.0 51.0 16.3 27.5 29.4 29.2 L102 550S162-33 48 3.7 3.7 3 16 58.6 31.1 28.0 17.5 21.3 20.6 20.8 L103 600S162-33 48 4.0 4.0 3 16 58.6 35.0 31.5 16.8 21.6 21.6 20.9 L104 1000S200-54 48 6.2 6.2 3 16 58.6 132.1 118.9 63.2 72.6 88.8 85 L105 800S162-43 48 5.0 5.0 3 16 58.6 67.8 61.1 32.2 37.0 43.8 42.7 L106 800S200-43 48 5.2 5.2 3 16 58.6 77.6 69.9 35.9 47.1 48.6 48 L107 600S200-33 48 4.2 4.2 3 16 58.6 40.5 36.5 18.4 24.8 NaN NaN L108 1200S200-54 48 7.2 7.2 3 16 58.6 172.4 155.2 58.1 69.5 108 107 L109 1000S200-43 48 6.2 6.2 3 16 58.6 106.4 95.8 32.3 46.4 62.2 62.1 L110 1000S250-43 48 6.5 6.5 3 16 58.6 119.5 107.6 35.2 47.7 63.4 61.6 L111 800S137-33 48 4.8 4.8 3 16 58.6 46.8 42.1 13.3 15.8 27.1 26.9 L112 800S162-33 48 5.0 5.0 3 16 58.6 52.4 47.2 14.7 21.6 29.3 29.8 L113 800S200-33 48 5.2 5.2 3 16 58.6 60.0 54.0 16.3 27.9 30.1 30.4 L114 550S162-33 48 2.9 2.9 3 16 58.6 31.1 29.5 16.1 21.8 22.3 22.2 L115 600S162-33 48 3.1 3.1 3 16 58.6 35.0 33.2 15.0 22.1 22.4 22.5 L116 1000S200-54 48 4.9 4.9 3 16 58.6 132.1 125.5 50.0 75.5 92.4 87.5 L117 800S162-43 48 3.9 3.9 3 16 58.6 67.8 64.4 25.6 38.5 46.2 43.9 L118 800S200-43 48 4.1 4.1 3 16 58.6 77.6 73.7 30.3 48.5 49.9 46.6 L119 600S200-33 48 3.3 3.3 3 16 58.6 40.5 38.5 17.8 25.3 23.6 23.3 L120 1200S200-54 48 5.7 5.7 3 16 58.6 172.4 163.8 44.3 72.7 113 113 L121 1000S200-43 48 4.9 4.9 3 16 58.6 106.4 101.1 25.6 48.0 64.5 63.4 L122 1000S250-43 48 5.1 5.1 3 16 58.6 119.5 113.5 29.3 49.1 64.4 63 L123 800S137-33 48 3.8 3.8 3 16 58.6 46.8 44.5 10.4 16.6 28.2 27.8 L124 800S162-33 48 4.0 4.0 3 16 58.6 52.4 49.8 12.0 22.4 30.9 30.3 L125 800S200-33 48 4.1 4.1 3 16 58.6 60.0 57.0 14.2 28.6 31.7 31 L
531
Appendix M 21BComparison of tested strengths to AISI S100-07 predicted strengths
The tested strengths of cold‐formed steel columns and beams with holes are
compared to AISI‐S100‐07 Main Specification strength predictions in this appendix. The
Main Specification test‐to‐predicted statistics are useful as a baseline comparison to the
DSM approaches developed in 1861HChapter 8. Main Specification test‐to‐predicted ratios are
obtained for the experimental data ( 1862HTable 4.3 and 1863HTable 4.5 for columns, 1864HTable 4.11 for
beams) and the simulated tests ( 1865HAppendix K for columns, 1866HAppendix L for beams)
described in this report. Member strengths are calculated with custom Matlab code
written by the author. The Matlab code implements Section C4 of AISI‐S100‐07 to
predict the ultimate strength of columns with holes and Section C3 of AISI‐S100‐07 to
predict the strength of laterally‐braced beams.
AISI‐S100‐07 considers two limit states for cold‐formed steel columns, (1) local‐
global buckling interaction (Section C4.1 of AISI‐S100‐07) and (2) distortional buckling
(Section C4.2 of AISI‐S100‐07). Column strength predictions for the local‐global
buckling limit state are calculated with the equation:
nen FAP =
532
where Ae is the effective area and Fn is the global column strength (stress). Ae includes
the influence of local buckling at the hole with the unstiffened strip approach for non‐
circular holes (see 1867HChapter 3 of this report) and employs empirical equations (Ortiz‐
Colberg 1981) for circular holes. Fn is determined with an empirical relationship
between global slenderness, λc, to column strength, where
e
yc F
F=λ
and Fy is the steel yield stress. The influence of holes on λc, i.e., the influence of holes on
the critical global elastic buckling stress, Fe, is not taken into account in the Main
Specification. (This is a fundamental difference of the DSM approach presented in this
report and the Main Specification.) Ae is indirectly capped at the net cross‐sectional area
of the section, Anet, with geometric limits in the Main Specification for both circular and
non‐circular holes. The net section cap is consistent with the DSM approaches in this
report. Pn for the distortional buckling limit state in the Main Specification is calculated
with the AISI‐S100‐07 Appendix 1 DSM approach. Currently the Main Specification
does not provide a method to account for the influence of holes on the distortional
buckling strength Pnd. The strength of the column is taken as the minimum of Pn (local‐
global interaction) and Pnd (distortional buckling).
For laterally‐braced beams, AISI‐S100‐07 considers two limit states, (1) local buckling
and (2) distortional buckling. Flexural strength predictions for the local buckling limit
state are calculated with the equation:
533
yen FSM =
where Se is the effective section modulus derived using the effective width method (see
AISI‐S100‐07 Section C3.1). A method is provided in the Main Specification to capture
the influence of a hole in the web of a beam on the local buckling strength with the
unstiffened strip approach (see AISI‐S100‐07 Section B2.4). The distortional buckling
strength, Mnd, is calculated with the DSM equations provided in AISI‐S100‐07 Appendix
1. As in the case for compression members, the Main Specification does not currently
provide a method to account for the presence of holes on Mnd. The strength of a beam is
taken as the minimum of Mn (local buckling) and Mnd (distortional buckling).
1868HTable M.1 through 1869HTable M.2 summarize the test‐to‐predicted statistics for the
column experiments and simulations. The statistics are presented to evaluate separately
those specimens with holes that lie within code limits and the specimens outside the
code limits (AISI‐S100‐07 provides the geometric limits for stiffened elements with holes
in Section B2.3, for stiffened element with holes under a stress gradient in Section B2.4,
note that there are no code limits for distortional buckling controlled specimens with
holes). The majority of the experiments are predicted to have a local‐global interaction
type failure (63 out of the 78 specimens). The test‐to‐predicted mean is 1.20 for the
experiments within code limits and 1.06 for specimens outside the code limits,
suggesting that column strength is sensitive to the geometric parameters considered in
the code. This trend is not observed in the simulation test‐to‐predicted mean, where the
mean is 1.04 for specimens within code limits and 1.07 for the specimens outside code
534
limits. The simulations tend to have larger, more closely spaced holes, resulting in a
test‐to‐predicted mean of 0.91 for distortional buckling controlled specimens. It is
hypothesized that the specimens designated as distortional buckling failures may
indeed be failing by local‐global interaction if the Main Specification is overpredicting
the local‐global interaction strength. This low test‐to‐predicted ratio highlights the
limitations of the current design approaches in AISI‐S100‐07 and confirms that the DSM
approaches developed in this report (see 1870HTable 8.1) are more viable predictors of column
strength when for hole geometries outside the current code limits.
The Main Specification is most accurate for stub columns as shown in 1871HTable M.3, and
is becomes more conservative for intermediate and long columns (i.e., as λc increases) in
1872HTable M.3. The increasly conservative strength prediction trend with increasing global
slenderness can be observed in 1873HFigure M.1. 1874HFigure M.2 demonstrates a similar trend for
the simulated column tests, where the Main Specification prediction becomes
unconservative with decreasing λc. The large distribution of distortional buckling
controlled‐specimens local‐global buckling interaction and the code is just missing the
behavior, especially because 212 of the 236 specimens are outside the Main
Specification’s geometric limits. 1875HFigure M.3 through 1876HFigure M.6 demonstrate that limits
imposed in the Main Specification on hole size and hole spacing, i.e. dh≤2.5 inches, Lh≤4.5
inches, S≥24 inches, and Send≥10 inches, are weak indicators of prediction viability and
that the dimensionless quantities studied in the element and member elastic buckling
535
studies in 1877HChapter 3 and 1878HChapter 4 of this report, , i.e., S/Lhole , S/h, hhole/h, may be more
viable as geometric design limits.
Table M.1 AISI S100‐07 test‐to‐predicted statistics for all column data with holes
Mean SD φ # of tests Mean SD φ # of testsall data 1.12 0.12 0.87 63 1.11 0.05 0.91 15within code limits 1.20 0.12 0.88 27outside code limits 1.06 0.06 0.91 36all data 1.06 0.13 0.86 236 0.91* 0.08 0.88 149within code limits 1.04 0.10 0.88 24outside code limits 1.07 0.13 0.86 212
* low test‐to‐predicted ratio results from L‐G buckling and D buckling predictions which do not adequately capture the influence of the hole
Shaded area means that code limits on hole geometry are not provided for the distortional buckling limit state in AISI‐S100‐07
Data Source Data Range
Experiments
Simulations
Local‐Global Buckling Interaction Distortional Buckling
Table M.2 AISI S100‐07 test‐to‐predicted statistics for stub columns with holes (λc≤0.20)
Mean SD φ # of tests Mean SD φ # of testsall data 1.07 0.07 0.90 38 1.03 ‐‐‐ ‐‐‐ 1within code limits 1.12 0.05 0.91 9outside code limits 1.05 0.07 0.90 29
Data Source Data Range
Experiments
Local‐Global Buckling Interaction Distortional Buckling
Table M.3 AISI S100‐07 test‐to‐predicted statistics for intermediate and long columns with holes (λc>0.20)
Mean SD φ # of tests Mean SD φ # of testsall data 1.20 0.13 0.87 25 1.11 0.05 0.91 14within code limits 1.24 0.14 0.87 18outside code limits 1.10 0.04 0.92 7all data 1.06 0.13 0.86 236 0.91* 0.08 0.88 149within code limits 1.04 0.10 0.88 24outside code limits 1.07 0.13 0.86 212
* low test‐to‐predicted ratio results from L‐G buckling and D buckling predictions which do not adequately capture the influence of the hole
Shaded area means that code limits on hole geometry are not provided for the distortional buckling limit state in AISI‐S100‐07
Experiments
Simulations
Data Source Data RangeLocal‐Global Buckling Interaction Distortional Buckling
536
0 0.5 1 1.5 2 2.5 30
0.2
0.4
0.6
0.8
1
1.2
1.4
1.6
1.8
2
λc=(Fy/Fe)0.5
Pte
st/P
n
AISI-S100-07 predicts local-global buckling limit stateAISI-S100-07 predicts dist. buckling limit state
Figure M.1 AISI S100‐07 test‐to‐predicted ratio for experiments on columns with holes, predictions become
increasingly conservative with increasing global slenderness
0 0.5 1 1.5 2 2.5 30
0.2
0.4
0.6
0.8
1
1.2
1.4
1.6
1.8
2
λc=(Fy/Fe)0.5
Pte
st/P
n
AISI-S100-07 predicts local-global buckling limit stateAISI-S100-07 predicts dist. buckling limit state
Figure M.2 AISI S100‐07 test‐to‐predicted ratio for FE simulated tests of columns with holes
537
0 1 2 3 4 5 6 7 80
0.2
0.4
0.6
0.8
1
1.2
1.4
1.6
1.8
2
dh, inches
Pte
st/P
n
AISI-S100-07 predicts local-global buckling limit stateAISI-S100-07 predicts dist. buckling limit state
Figure M.3 AISI S100‐07 test‐to‐predicted ratio, FE simulated tests of columns with holes, code requires that
hole depth dh≤2.5 inches
0 1 2 3 4 5 6 7 80
0.2
0.4
0.6
0.8
1
1.2
1.4
1.6
1.8
2
Lh, inches
Pte
st/P
n
AISI-S100-07 predicts local-global buckling limit stateAISI-S100-07 predicts dist. buckling limit state
Figure M.4 AISI S100‐07 test‐to‐predicted ratio, FE simulated tests of columns with holes, code requires that
hole length Lh≤4.5 inches
538
0 5 10 15 20 25 300
0.2
0.4
0.6
0.8
1
1.2
1.4
1.6
1.8
2
S, inches
Pte
st/P
n
AISI-S100-07 predicts local-global buckling limit stateAISI-S100-07 predicts dist. buckling limit state
Figure M.5 AISI S100‐07 test‐to‐predicted ratio, FE simulated tests of columns with holes, code requires hole
spacing S≥24 inches
0 5 10 15 20 25 300
0.2
0.4
0.6
0.8
1
1.2
1.4
1.6
1.8
2
Send, inches
Pte
st/P
n
AISI-S100-07 predicts local-global buckling limit stateAISI-S100-07 predicts dist. buckling limit state
Figure M.6 AISI S100‐07 test‐to‐predicted ratio, FE simulated tests of columns with holes, code requires
clear hole spacing at column ends Send≥10 inches
The Main Specification test‐to‐predicted comparison for the laterally‐braced beam
experiments and simulated tests are summarized in 1879HTable M.4. The test‐to‐predicted
539
mean is low for all experiments, which is consistent with the DSM comparison in 1880HTable
8.5. It is still unclear what why the tested strengths are systematically lower than the
predictions, and an investigation is ongoing. Conclusions can still be drawn from these
results though. It is observed that the experiments with hole geometries within code
limits are predicted better than beams outside of code limits (mean test‐to‐predicted
ratio of 0.90 versus 0.80), suggesting that the viability of the Main Specification
prediction method is related to the geometric limits. The simulated beam tests also
exhibit a low test‐to‐predicted ratio of 0.92 for the local‐global buckling interaction
specimens, but in this case for a different reason than the experients. It is hypothesized
that the Main Specification is not as viable when the hole geometries are outside code
limits. When comparing the results in 1881HTable M.4 with the test‐to‐predicted statistics for
the DSM approaches in 1882HTable 8.4, it becomes evident that DSM (for beams with holes) is
a more viable predictor of strength over a wider range of hole geometries.
Table M.4 AISI S100‐07 test‐to‐predicted statistics for laterally braced beams with holes
Mean SD φ # of tests Mean SD φ # of testsall data 0.83 0.12 0.84 64 0.90 0.12 0.85 80within code limits 0.90 0.16 0.80 21outside code limits 0.80 0.07 0.88 43all data 0.92* 0.07 0.89 20 1.07 0.11 0.87 184within code limits ‐‐‐ ‐‐‐ ‐‐‐ 0outside code limits 0.92* 0.07 0.89 20
* low test‐to‐predicted ratio results from L‐G buckling and D buckling predictions which do not adequately capture the influence of the hole
** Shan and LaBoube beam strengths are low compared to predictions even for members without holes, Mean<1 does not necessarily suggest a hole influence
Experiments**
Simulations
Data Source Data RangeLocal‐Global Buckling Interaction Distortional Buckling
Re
sea
rch
Re
po
rt R
P-0
9-1
American Iron and Steel Institute
1140 Connecticut Avenue, NW Suite 705 Washington, DC 20036 www.steel.org